--- category: literaturenote citekey: thomasdetailingaccessibilitydimensiontransport title: "Detailing the accessibility dimension of transport poverty: an application in Tarragona" authors: "Thomas, Kilian" zotero_key: HZKFDW7U zotero_storage: 9KP85GK2 collections: imporditud folder: 001_artiklid firstAuthor: "Thomas, Kilian" status: converted --- ![](_page_0_Picture_0.jpeg) # MASTER Detailing the accessibility dimension of transport poverty an application in Tarragona | Thomas, Kilian C.F. | | |---------------------|--| | Award date:
2025 | | [Link to publication](https://research.tue.nl/en/studentTheses/dfccf742-f31a-4137-9191-77575e465be0) #### Disclaimer This document contains a student thesis (bachelor's or master's), as authored by a student at Eindhoven University of Technology. Student theses are made available in the TU/e repository upon obtaining the required degree. The grade received is not published on the document as presented in the repository. The required complexity or quality of research of student theses may vary by program, and the required minimum study period may vary in duration. #### General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. - Users may download and print one copy of any publication from the public portal for the purpose of private study or research. - You may not further distribute the material or use it for any profit-making activity or commercial gain #### Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Download date: 06. mai. 2026 ![](_page_1_Picture_0.jpeg) # Detailing the accessibility dimension of transport poverty: an application in Tarragona Kilian Thomas Student ID: 2219263 Surname and initials: Thomas, K.C.F. # Graduation Supervision Committee: dr.ing. Peter J.H.J. van der Waerden Javier Mart´ınez Boada dr. Feixiong Liao Master thesis, July 2025 Master ABP, track Sustainable Urban Mobility Transitions Course code: 7SU30M0 - 30 EC #### Disclaimer: This thesis has been carried out in accordance with the rules of the TU/e Code of Scientific Integrity. This graduation thesis is publicly available. # General Title: Detailing the accessibility dimension of transport poverty: an application in Tar- ragona. Date: July 2025 Author: Thomas, Kilian Christophe Fr´ed´eric (K.C.F.) Student number: 2219263 Department: Department of the Built Environment Master program: Architecture, Building and Planning Master track: Sustainable Urban Mobility Transitions (SUMT) Part of: EIT Urban Mobility Master School Course code : Graduation project sustainable urban mobility transitions (7SU30M0) EC: 30 EC Evaluation: The weights of each part relative to the assessment should be the ones set by the master track Company : Ingenier´ıa y Econom´ıa del Transporte S.M.E. M.P., S.A. (INECO) ### Supervision committee Main supervisor: dr.ing. Peter J.H.J. van der Waerden from Department of the Built En- vironment, TU/e Second supervisor and chair: dr. Feixiong Liao from Department of the Built Environ- ment, TU/e Third supervisor: Javier Mart´ınez Boada from INECO ### Disclaimer This thesis has been carried out in accordance with the rules of the TU/e Code of Scientific Integrity. This graduation thesis is publicly available. # Preface I want to extend my heartfelt thanks to everyone who has supported and helped me throughout this journey. Thank you, Peter, for your guidance and insightful comments during our meetings. Your questions have greatly enriched this work. I am also deeply grateful to Javier, for his valuable insights and for always keeping me updated on the latest publications in the field. Also for proposing this subject, which was unknown to me before starting this thesis, and which I found very interesting throughout the entire investigation process. Finally, to Professor Liao for his pertinent suggestions. I also want to thank my parents, for their constant support and care throughout my master's program. Your help has been invaluable. Ana, thank you for always being there and for your patience. Lastly, I want to express my gratitude to all my friends from the master's program. I hope that life will bring us together again soon. # Executive summary Principle 20 of the European Pillar of Social Rights states that everyone has the right to access essential services. Transport is crucial in this regard, as it enables people to reach the facilities that provide these services. In addition, adequate transport is relevant for a region competitiveness, plays a vital role in promoting economic development and contributes to enhancing intergenerational fairness and solidarity. However, not all people have the same opportunities and facilities to access and use transport, and for those less fortunate, this could translate into transport poverty. Transport poverty is a term that encapsulates all problems related to transport people may endure, and that do not enable them to actively and fully take part in society. These problems range from lack of available transport options, to inadequate (non-adapted) vehicles, or to lack of financial means. As for the causes of transport poverty, they are broad, and can stem from factors such as low income but also from other systemic barriers, such as lack of access to essential facilities, geographic isolation or limited availability of public transport. As can be seen transport poverty encompasses a wide variety of problems, and for this reason, this concept is usually subdivided into smaller pieces, that are called dimensions, and that facilitate its study. The four transport poverty dimensions are: availability, accessibility, affordability and adequacy. All of them require more research, as this concept is very broad and covers numerous areas of study. Yet, accessibility (as well as adequacy) are the ones requiring more attention, as pointed out in a report commissioned by the Spanish government about transport poverty published in 2025. A similar finding is presented in the report commissioned by the EU to investigate transport poverty and published in 2024, that only offers one indicator for the accessibility dimensions, while several are proposed for other dimensions such as availability or affordability. Therefore, this study focuses on one of them: accessibility. In particular, the main research question of this thesis is how the accessibility dimension of transport poverty can be measured and presented, so accessibility measurements become useful for decision-makers when allocating transport investments. In addition, this thesis also aims to unravel how accessibility is linked to transport poverty, what data is needed to measure accessibility in this context and how could accessibility in rural areas be assessed, as these areas have been identified as particularly at risk from suffering from transport poverty. Addressing this last point is particularly relevant since the analysis of transport in rural areas is significantly different from the one in urban areas. This is mainly due to two reasons. First, the fact that rural areas do not have enough demand to host all the essential facilities its residents need, so their inhabitants have to travel outside these areas to access them. This is often not the case in urban areas, where trips to essential facilities take place inside them. The second reason is the different transport network. Usually more options are available in urban cities (metro, tram, etc.), while in rural areas the choice is much more restricted. To answer all these questions, the main output of this thesis is the creation of a methodology, that builds upon and refines the methodology proposed in a recent publication in the field, to measure accessibility in the context of transport poverty. This measurement is grounded on the accessibility definition offered by the European Commission, and that computes accessibility as the percentage of essential trips that can be done within a reasonable amount of time. To that end, several parameters need to be established. In the first place, it is required to determine the desired level of geometric granularity for the study (province, municipality, neighbourhood, etc.), and to classify the population in several groups of interest (youngsters, elderly, impaired, etc.). The second step is to establish, for each of these groups, which are the essential services they need to have access to (namely which are the facilities providing access to them), how frequently should they have access to them (for example how many times a month), and what is the maximum time that reaching these locations should take. Having set all these parameters, it is possible to determine how many essential trips should be done per month for each person in each group of interest. After that, the proposed methodology requires to compute the actual travel times it takes the citizens in each area of study to reach those previously defined essential facilities, and to evaluate if they are above or below the previously defined time thresholds. The proposed computation methods for travel time focuses in public transport, aligning with the Spanish ambitions and the European directives that encourage public transport and promote sustainable ways of transport. To compute travel times, network information is required, and for the case of public transport namely GTFS data, which is open source. With this method, it is possible to compute the percentage of essential trips that can be done in a reasonable amount of time, this is accessibility. It is noted that because of only considering the trips providing access to essential destinations, this indicator is suited to assess accessibility in the context of transport poverty. Not being able to carry any of these trips means that individuals will not be able to fully participate in society, and therefore will suffer from transport poverty. To prove the applicability of the proposed methodology, a case study in Tarragona is presented. The objective of it is to showcase how this methodology can output relevant insights in the field of accessibility in the context of transport poverty. Although simplifications are applied to reduce computation time, it is possible to assess the promising results that it could offer. Applying this method, it is possible to gain insides on which regions have or do not have access to each essential facility, and therefore have enough accessibility not to suffer from transport poverty. The information provided further enables for the identification of which regions are more severely affected by low accessibility, providing guidance when prioritizing the interventions. For instance, it is possible to identify that out of the 125 rural municipalities in Tarragona, 107 suffer from transport poverty (85.6 %), or that for 64 municipalities, travel times to hospitals are too long. Improving access to regional heads could greatly improve this transport poverty situation. The case study also provides insights on how the choices made when applying the methodology can output results that focus more on sustainable alternatives such as public transport, or in vulnerable groups such as inhabitants of rural areas, making the results aligned and coherent with the European and Spanish interests. This measure also proves informative to compare accessibility levels between different groups with different needs, identifying that in Tarragona, accessibility by public transport for adults is lower than for youngsters and elderly. Lastly, the measurement offered is further refined to not only consider the number of trips that can not be done, but also to account for the number of citizens that are affected by this situation, among other possible improvements. Several strengths and weaknesses are identified. On the one hand, this methodology is capable of answering the research question, since it allows to quantify accessibility in the context of transport poverty, and provide decision makers with objective, visual, comprehensible and data-driven indicators to prioritize interventions. It also helps understanding which is the importance of accessibility dimension in the context of transport poverty, and how different vulnerable groups needs can be emphasized. Another asset of this methodology is that it is based in currently available data and that it is replicable in other countries and that results obtained are comparable across groups and regions. However, limitations are also found. In the first place, the concept of transport poverty is still evolving and therefore the notion of essential trips and reasonable amount of time are not fixed yet, both being crucial information for this indicator. Apart from that, this method requires big amounts of data which may not be available universally, and even when they are, is computationally intensive. Lastly, this indicator does not take into account the personal decisions and choices, an important element when studying transport. All in all, this thesis presents valuable insights. On the one hand, by framing and detailing the concept of accessibility in the context of transport poverty. On the other hand, it offers an intuitive and visual indicator that quickly identifies areas suffering from a lack of accessibility (and thus from transport poverty), precisely indicating the unmet needs. # Contents | | | Executive summary | iv | |---|--------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|------------------------------------------------| | | Contents | | vii | | | | List of Figures | x | | | | List of Tables | xii | | 1 | 1.1
1.2
1.3
1.4 | Introduction
Introduction: transport poverty

1.1.1
Transport poverty, a consequence of the need to travel

1.1.2
A European problem, intensified in rural areas

1.1.3
Transport poverty in the Spanish context

Problem statement

Research question and sub-questions

Academic and societal relevance

1.4.1
Academic relevance

1.4.2
Societal relevance
| 1
1
1
2
3
4
6
6
6
7 | | | 1.5 | Research design
| 8 | | 2 | 2.1 | Literature review
Definitions

2.1.1
Transport poverty

2.1.2
Accessibility in the context of transport poverty

2.1.3
Rural areas
| 10
10
10
13
15 | | | 2.2 | Measuring accessibility in transport

2.2.1
Currently used indicators

2.2.1.1
Accessibility based on travel time to key destinations

2.2.1.2
Accessibility as a mix of transport related indicators

2.2.1.3
Accessibility based on access to essential facilities

2.2.2
Computing average travel time to decentralized services
| 15
15
16
19
20
22 | | | 2.3 | Decision making in transport

2.3.1
Particularities of the decision making in transport

2.3.2
Main political strategies related to the topic

2.3.2.1
Global and European policies

2.3.2.2
Spanish Policies

2.3.2.3
Conclusion of the main directives
| 25
26
27
27
29
30 | CONTENTS CONTENTS | | 2.4 | Concluding remarks

30 | | |---|-----|-----------------------------------------------------------------------------------------------|--| | 3 | | Methodology
32 | | | | 3.1 | Link between accessibility and transport poverty

32 | | | | 3.2 | Proposal of an accessibility indicator

36 | | | | | 3.2.1
Exposition of the suggested indicator

36 | | | | | 3.2.2
Required data for the accessibility indicator

40 | | | | | 3.2.2.1
List of essential facilities, frequencies and acceptable travel times
40 | | | | | 3.2.2.2
Other raw data required

42 | | | | | 3.2.3
Calculation of the travel times and its consequences on the indicator
43 | | | | | 3.2.3.1
Computation method of travel times for public transport

43 | | | | | 3.2.3.2
Implications of the computation method

45 | | | | | 3.2.4
Assumptions and motivations of the indicator

46 | | | | | 3.2.5
Comparison of the proposed methodology with Radics et al. (2024).
47 | | | | 3.3 | Conclusions

48 | | | | | | | | 4 | 4.1 | Results. A case study in Tarragona
50
Tarragona, a region in Catalu˜na, Spain

50 | | | | 4.2 | Establishment of the different parameters required for the measurement

53 | | | | | 4.2.1
Definition of the target groups

53 | | | | | 4.2.2
Definition of essential facilities

54 | | | | | 4.2.3
Definition of frequency of essential trips

56 | | | | | 4.2.4
Definition of other parameters

58 | | | | | 4.2.5
Implications: a public transport indicator for rural areas

60 | | | | 4.3 | Intermediate results: results per category

61 | | | | | 4.3.1
Results for healthcare

61 | | | | | 4.3.2
Results for work

64 | | | | | 4.3.3
Results for education

64 | | | | | 4.3.4
Results for commerce

65 | | | | | 4.3.5
Results for leisure

65 | | | | 4.4 | Final accessibility results

67 | | | | | 4.4.1
Accessibility general results

67 | | | | | 4.4.2
Accessibility for youngsters, adults and elderly

69 | | | | | 4.4.3
Policy recommendations

71 | | | | 4.5 | Conclusions

73 | | | 5 | | Discussion
75 | | | | 5.1 | Strengths and limitations of the methodology

75 | | | | | 5.1.1
Strengths of methodology

75 | | | | | 5.1.2
Limitations of methodology

76 | | | | 5.2 | Further improvements of the methodology

78 | | | | | 5.2.1
Expanding the indicator including number of citizens

78 | | | | | 5.2.2
Expanding the indicator modifying compliance requirements

81 | | | | | 5.2.3
Expanding the methodology using catchment areas for each service

82 | | | | | 5.2.4
Other improvements

83 | | | | 5.3 | Applicability in other contexts, a case study of the Netherlands

85 | | | | | 5.3.1
Administrative units and population disaggregation

85 | | CONTENTS CONTENTS | Appendix | | | 104 | |----------|------------|---------------------------------------------------------------------------|-----| | 5.5 | Conclusion | | 94 | | | 5.4.2 | Assessing municipalities connectivity to their regional heads
| 90 | | | 5.4.1 | Comparison with car-based accessibility measures
| 89 | | 5.4 | | Other results of interest
| 89 | | | 5.3.5 | Transport poverty in the Netherlands, a relevant issue
| 88 | | | 5.3.4 | Data availability and model transferability
| 87 | | | 5.3.3 | Time thresholds adapted for a more urban context
| 86 | | | 5.3.2 | Essential facilities selection and setting of essential trips frequencies | 85 | # List of Figures | 1.1 | Research design of the thesis.
| 9 | |-----|------------------------------------------------------------------------------------------------|----| | 2.1 | Transport poverty dimensions, based on Cludius et al. (2024).
| 13 | | 3.1 | Flowchart to evaluate if an individual is suffering from transport poverty, a | | | | proposal.
| 33 | | 3.2 | Illustrative application of the flowchart.
| 34 | | 3.3 | Illustration of the methodology for computing travel times.
| 45 | | 4.1 | Names of several municipalities of Tarragona Province.
| 52 | | 4.2 | Population per municipality in Tarragona according to
INE Census 2024
data. | 53 | | 4.3 | Distribution of the monthly essential trips and total number of trips per group | | | | based on the established parameters.
| 58 | | 4.4 | Accessibility by bus to hospitals.
| 62 | | 4.5 | Accessibility by bus to primary medical attention centres (CAPs).
| 63 | | 4.6 | Accessibility by bus to municipal sports centres.
| 63 | | 4.7 | Accessibility by bus to regional heads (Cabezas comarcales), a proxy for work. | 64 | | 4.8 | Accessibility by bus to high schools.
| 65 | | 4.9 | Accessibility by bus to supermarkets.
| 66 | | | 4.10 Accessibility by bus to shopping malls.
| 66 | | | 4.11 Accessibility for all municipalities.
| 69 | | | 4.12 Accessibility per municipality in rural areas.
| 70 | | | 4.13 Accessibility per municipality for different age groups.
| 71 | | | 4.14 Accessibility per municipality for different age groups in rural areas.
| 71 | | 5.1 | Comparison of unfeasible trips per municipality and per citizen.
| 80 | | 5.2 | Comparison of unfeasible trips per municipality and per citizen in rural areas. | 81 | | 5.3 | Line R16 of
Rodalies
going through Tarragona [Rodalies de Catalunya].
| 82 | | 5.4 | Excess time spent in public transport per municipality and per person.
| 82 | | 5.5 | Accessibility by bus to hospitals arriving before 8:30 AM.
| 83 | | 5.6 | Accessibility by bus to hospitals only considering municipalities with more than | | | | one expedition per day.
| 84 | | 5.7 | Illustration of the methodology for computing travel times accounting for catch | | | | ment areas.
| 84 | | 5.8 | Urban-rural typology according to EU classification.
Predominantly urban | | | | regions (blue), intermediate regions (brown), rural areas (green).
[Eurostat,
2025].
| 87 | | | | | LIST OF FIGURES LIST OF FIGURES | 5.9 | Classification of the degree of urbanization according to the
CBS classification. | | |-----|----------------------------------------------------------------------------------------|----| | | Red:
extremely urbanised, blue:
strongly urbanised, light green:
moderately | | | | urbanised, dark green: hardly urbanised , pink: not urbanized (equivalent to | | | | rural areas).
| 87 | | | 5.10 Comparison of accessibility by car and public transport to cities with more | | | | than 50,000 inhabitants in 45 minutes according to
Eursotats
(car, left),
IGN | | | | (car, center) and the methodology of this thesis (bus, right).
| 90 | | | 5.11 Comparison of accessibility to the closest 5,000 inhabitants municipalities by | | | | car (left) or bus (right).
| 91 | | | 5.12 Comparison of accessibility to the closest 20,000 inhabitants municipalities by | | | | car (left) or bus (right).
| 92 | | | 5.13 Comparison of accessibility to the closest 50,000 inhabitants municipalities by | | | | car (left) or bus (right).
| 92 | | | 5.14 Comparison between better connected regional head and assigned administrat | | | | ive regional head using bus per municipality.
| 93 | # List of Tables | 2.1 | Definitions of transport poverty from various sources.
| 11 | |-----|------------------------------------------------------------------------------------------------------------------------------------------|-----| | 3.1 | Proposed indicator to measure accessibility in a municipality
m
for
K
groups
and
L
essential facilities.
| 39 | | 4.1 | Table of essential facilities and the sources of their location data for the case
study.
| 55 | | 4.2 | Frequency of visits per month and per group to each essential facility for case
study.
| 58 | | 4.3 | Complete accessibility table for
Ulldemolins
municipality.
| 67 | | 4.4 | Travel times for the municipality of
L'Argentera.
| 70 | | 5.1 | Comparison of travel times by bus and car to the closest municipalities with
5,000, 20,000 and 50,000 inhabitants municipalities.
| 90 | | 2 | Transport availability by income level and population group. Sample Table for
one area of study in the illustrative example.
| 106 | | 3 | Results of accessibility analysis for the example for a village under study, cit | | | | izens with
<2,000
€/month. An illustrative example.
| 108 | | 4 | Accessibility results for the population of a village under study. An illustrative
example.
| 109 | # Chapter 1 # Introduction # 1.1 Introduction: transport poverty ### 1.1.1 Transport poverty, a consequence of the need to travel The historical transformation of cities to vast, centralized, and interdependent urban hubs—has significantly influenced societal behaviour and created new challenges. As stated in McPhearson et al. [\(2016\)](#page-113-0) "Cities and urbanized regions are complex, dynamic, and highly integrated systems (...) that create deep challenges for good governance, policymaking, and planning" (p.1). Modern cities are not only far more complex than those of previous centuries, but have also undergone a shift toward the centralization of essential services. The establishment of hospitals, universities, and business districts has concentrated the provision of services in specific areas, necessitating physical presence for access. Consequently, travel has become an unavoidable part for people wanting to access those services. In some countries, such as the United States, urban planning has further amplified this need by relocating residential areas to suburban outskirts, increasing even more reliance on transportation to participate in society (Council of Europe, [2008\)](#page-110-0). Thus, urban development and spatial planning has shaped both residential patterns, but also travel behaviour as individuals navigate the built environment (Farinloye et al., [2019,](#page-111-1) Guagliardo, [2004,](#page-112-0) Morency et al., [2011\)](#page-114-1). However, the necessity to travel is not solely dictated by the physical distribution of services; personal preferences also play a crucial role. Beyond the spatial organization of cities, factors such as individual attitudes, lifestyle, perceptions, preferences and socio-demographics influence travel choices (Handy et al., [2005,](#page-112-1) Van Acker et al., [2010\)](#page-115-0). This translates into the fact that, even when facilities providing essential services are available nearby, individuals may opt for others that may be further away based on subjective factors. The study of travel behaviour and travel patterns seeks to understand these complex decision-making processes. Numerous publications explore why people travel and how they make transportation-related decisions (Van Acker et al., [2010,](#page-115-0) Buliung and Kanaroglou, [2007,](#page-109-1) Hanson and Hanson, [1981,](#page-112-2) Etminani-Ghasrodashti and Ardeshiri, [2015\)](#page-110-1). Three key conclusions can be drawn from this discussion. First, travel has become a fundamental requirement for accessing certain services, with little to no alternative in many cases. Second, the way individuals decide on how to travel is shaped by two main factors: spatial planning and personal circumstances. On the one hand, spatial or urban planning determines the location of essential facilities and the infrastructure that makes them accessible. On the other hand, personal preferences, socio-economic conditions, and education influence individual travel choices. While the former relates to strategic planning and highlevel decision-making, and is often related in the case of transport to significant investments for infrastructure and vehicles, the latter is rooted in individual decision-making processes. Both matters are intrinsically linked. The decisions taken by the strategic planners influence the choices made by the users, modifying their preferences. And, at the same time, the preferences of the majority of the users are taken into account by their elected representatives, in charge of the decision making. In this thesis, and although both of them are relevant topics to understand transportation, the primarily focus will be on the first aspect: strategic decision-making in urban planning. The third conclusion appears as a consequence of the two previous ones. Because transport has become a necessity, yet its availability and convenience is influenced by strategic decisions and personal preferences, inequalities arise, leading to disparities in access to services. When accessibility is severely limited, this can translate or be categorized as transport poverty. In a recently published recommendation of the European Commission (European Commission, [2025\)](#page-111-2), it is stated that "causes of transport poverty stem from low income and other systemic barriers, such as lack of access to (...) educational facilities and essential services, geographic isolation, absence or limited availability of public or private transport or specific socio-economic, demographic and physical characteristics that limit individuals" (p.1). Therefore, it becomes relevant to analyse what is transport poverty, how to define it, understand what is causing it and more importantly, try to find effective ways to counter it. # 1.1.2 A European problem, intensified in rural areas The European Union is committed to achieving net zero emissions by 2050 (European Parliament and Council, [2019\)](#page-111-3). This implies to stop using non-renewables sources of energy such as fossil fuels, in favour of more sustainable alternatives, such as sun or wind power. This concept is broadly referred to as the sustainable transition. And within this framework, the transport sector plays a crucial role. Transport accounted for approximately 29% of the EU's greenhouse gas emissions in 2022 (European Environment Agency, [2025\)](#page-111-4), making it a key target for decarbonization efforts. However, while the transition to sustainability presents numerous opportunities, it also carries the risk of disproportionately affecting certain groups (Alguacil Denche et al., [2024\)](#page-109-2). Ensuring a just transition that leaves no one behind is at stage (European Commission, [2021c\)](#page-111-5). In this context, "transport poverty is increasingly a concern, in particular for vulnerable groups" (European Commission, [2025,](#page-111-2) p.1). The term of transport poverty encapsulates the notion that, because of not perfect transport, certain segments of the population experience unequal access to essential services, thereby limiting their full participation in society (Cludius et al., [2024\)](#page-109-0). Recognizing the significance of this issue, the European Parliament underscored in 2022 the necessity for further research, acknowledging that at that time "the concept of 'transport poverty' is recent and has no established definition in the academic or policy literature so far" (Kiss, [2022,](#page-113-1) p.1). As a result, researching this matter and refining transport poverty definition and implications became a priority for the EU. Consequently, EU institutions commissioned a report that delved into these issues in greater detail. The challenges of this study were to identify what was transport poverty, offer insights on each of the dimensions, and propose measurements to measure them across EU countries, in order to ensure equitable policy interventions, and optimize the allocation of EU social funds. The findings of this publicly available study, published in 2024 (see Cludius et al., [2024\)](#page-109-0), serve as a foundational reference for the present research, providing essential insights into the nature and implications of transport poverty. They are also aligned with the transport poverty definition established in the Social Climate Fund Regulation of 2023 (European Parliament and Council, [2024\)](#page-111-6). Although the study provided valuable insights, its broad scope—encompassing all European Union countries—prevented a detailed analysis of the most impacted subgroups. And notably, it did not include specific insights for rural communities, which are considered particularly vulnerable to transport poverty. Indeed, the European Parliament's resolution of 2022 on the subject already highlighted that "mobility poverty has been underexposed (...). However, it is a problem that is becoming more pressing (...) for those living in rural (...) areas" (European Parliament and Council, [2022,](#page-111-7) art.1, para.3). Further refining this perspective, Amendment 55 defined mobility poverty as a "condition affecting households within the lowest income deciles, including lower middle-income groups, with a particular emphasis on those residing in rural or less accessible areas" (art.1, para.1). However, focusing on rural areas has important implications when analysing transport networks, as they differ significantly from urban areas. For many essential facilities, such as schools, hospitals, or supermarkets, a minimum demand (number of users) is required to justify the expenses linked to creating and maintaining them. For instance, it is expected that in 2050, the Spanish student population will decrease by 800,000 people, the equivalent of 33,000 classes of 24 students. This will most likely translate, particularly in rural areas, to the closure of school facilities due to lack of demand (Gobierno de Espa˜na, [2021a\)](#page-111-8). Consequently, "rural populations have less access to services and activities (...)" (United Nations Economic Commission for Europe, [2017,](#page-115-1) p.1). This often translates to the fact that, while urban area citizens have to navigate within their cities to access services, rural populations often need to travel outside their settlements or municipalities, often to the closest urban area, in order to receive them. This notion is captured in the urban-rural catchment areas (URCAs) explored for instance in Cattaneo et al. [\(2021\)](#page-109-3)). The need to go outside their settlements to access services, translates in longer travel distances for rural areas inhabitants. Endorsing this statement, in their study Kompil et al. [\(2022\)](#page-113-2) expose how travel distances to access facilities such as retailers, schools or pharmacies are consistently higher for residents of rural areas than for cities inhabitants. Longer distances will lead to both; generally longer travel times, and the dismissal of some means of transport (mainly active mobility), which become unfeasible due to the high physical distance. In this sense, while urban citizens will have at their reach numerous alternatives, such as walking, cycling, driving or taking urban means of transport (urban buses, tramways or metros), citizens in rural areas will commonly be faced with only two options: private car or longdistance public transport such as trains or intermunicipality buses. Therefore, the network analysis for both areas is significantly different. ### 1.1.3 Transport poverty in the Spanish context The issue of transport poverty takes on particular importance in rural areas in Spain. Despite representing only 15.9% of the national population— still a very significant amount of 7,538,929 people as of 2020—Spain's rural areas cover an overwhelming 84% of the country's territory , asshowin in Gobierno de Espa˜na [\(2021c\)](#page-112-3). This disparity between population density and geographic coverage presents unique challenges, particularly regarding mobility and accessibility. The need to tackle this problem is not only pressing from a mobility point of view. It is also required to prevent the ongoing risk of depopulation. Alarmingly, in Spain up to 50% of rural municipalities face the threat of disappearing due to limited economic opportunities. Because of all the consequences this trend entangles (Gobierno de Espa˜na, [2021b\)](#page-112-4), this has become a very serious issue. In response, the Spanish government launched the "Demographical Challenge Campaign" in 2021 (Gobierno de Espa˜na, [2021b\)](#page-112-4), a strategic initiative aimed at countering rural decline and fostering sustainable development. Among all 130 measures proposed, several of them are specifically related to transport (ex. 1.18 or 2.7), emphasising that these two problems are tightly related. Therefore, a correct diagnosis of transport poverty situations will not only help ensuring social and economic inclusion of population living in these critical areas, but will also help counteract the rural exodus currently draining the population from Spanish rural areas. In addition, the Spanish government also has precise ambitions when it comes to mobility. In the Spanish mobility strategy, (Ministerio de Transportes y Movilidad Sostenible, [2021\)](#page-114-2), it is clearly stated that "Mobility should be considered as a right" (p.45), and that "Mobility should guarantee that citizens can cover their needs, without requiring their own vehicle" (p.62). There is therefore a strong will to put public transport in the spotlight. This positioning also aligns with the EU Green (European Parliament and Council, [2019\)](#page-111-3) and the fit for 55 measures (European Council, [2021\)](#page-111-9), that put in the centre of the debate the most sustainable alternatives. As shown in Ritchie [\(2023\)](#page-114-3), public transport has a lower footprint than private transport, even when compared to electric cars. However, the vast territory, the risk of depopulation and the Spanish mobility ambitions are not the only factors that make this a relevant problem with a complex solution. Spain's governance structure further complicates the implementation of transport-related policies. The constitutional framework grants autonomous communities significant competencies in territorial planning, public services, economic development, and tourism. As a result, mitigating transport poverty requires a shared commitment across multiple levels of government. Coordination among national, regional, and local authorities is essential, and cooperative governance mechanisms must be implemented effectively (Gobierno de Espa˜na, [2021b\)](#page-112-4). As a result, investigating transport poverty in rural areas in Spain is imperative. It is needed to develop a methodology to understand which areas are suffering more from this, to create indicators to measure and compare situations, and to provide decision makers with the required tools to efficiently allocate resources. This will not only create a more fair society, but will also increase the quality of life in rural areas, preventing depopulation, and will help the Spanish government pursue its ambitions. # 1.2 Problem statement Recognizing the urgency of the aforementioned issues, the government ministry of transportation launched the "Spanish Strategic Mobility Plan" (Ministerio de Transportes y Movilidad Sostenible, [2021\)](#page-114-2). This strategic set of measures establishes that transport should be regarded as a fundamental right, and therefore all Spanish citizens should have and gain accessibility from public transport. While this is applicable to the whole territory, it becomes a major issue in rural areas, that tend to be worse connected to the economic centres and to other essential facilities by public transport, leading to a bigger reliance on cars (OTLE, [2025\)](#page-114-4). Transportation is a vital link that connects people to their communities, jobs, and essential services, making it a key factor in social cohesion and inclusivity. When transport systems are inadequate, they can isolate individuals, preventing them from fully participating in daily life and contributing to the economy (European Commission, [2025\)](#page-111-2). While solutions like increasing public transport frequency and availability are effective and well understood, they often come with significant costs. This makes it crucial to prioritize resources effectively, ensuring that the most vulnerable citizens and regions receive the support they need first. Consequently, it is required to develop methods to assess the extent of transport deprivation in rural regions, enabling authorities to identify areas where investments are most urgently needed. This view aligns with the EU Mobility Strategy (European Commission, [2021c\)](#page-111-5), that states that "Fostering cohesion, reducing regional disparities as well as improving connectivity and access to the internal market for all regions, remains of strategic importance for the EU" (p.2). Spain is aware of the importance of this, and has already launched a study, published in April 2025, about Transport poverty (see OTLE, [2025\)](#page-114-4). In this document, it is recognized that "rural areas (...) usually have less availability and accessibility to public transport, becoming more reliant on cars" (p.15). As awareness of these challenges grows, the concept of Transport poverty has gained increasing recognition, offering a framework to better understand and address the issue. Over time, this framework has become more defined, yet key aspects still require further clarification to make the problem more tangible and manageable. To this day, transport poverty is typically assessed through four key dimensions: availability, adequacy, affordability, and accessibility(Cludius et al., [2024\)](#page-109-0), that will be thoroughly reviewed in Section [2.1.1.](#page-22-2) Among these, accessibility, directly related to travel time to destinations, stands out as the most pressing concern in Spain. As indicated in the transport poverty analysis commissioned by the Spanish government (OTLE, [2025\)](#page-114-4), "very few indicators have been developed to quantify transport poverty, particularly for accessibility and adequacy dimensions" (p.19). In part, this is because analysing this concept is particularly challenging. Accessibility is closely related to other dimensions of transport poverty (in particular availability), making it difficult to isolate it (see Section [2.1.1\)](#page-22-2). Additionally, improving accessibility is not as straightforward as addressing other dimensions of transport poverty. For instance, diminishing travel time to hospitals (improving accessibility to hospitals), could involve better transport connections to existing facilities, or the construction of new hospitals in underserved areas. This shows that the answers to poor accessibility may not be transport related (Metta, [2020\)](#page-113-3). Moreover, accessibility is inherently difficult to quantify, as it often relies on large-scale GIS datasets and often requires advanced modelling expertise. Given these complexities, this document places particular emphasis on accessibility and its implications. Beyond these technical and policy challenges, an additional complexity arises concerning accessibility: defining which destinations should be accessible, this is, what should be the list of essential facilities every citizen should be able to access in a reasonable amount of time. Establishing a clear set of locations that citizens need access to—as well as identifying the origins from which accessibility should be measured—is not a settled matter. There is also room for debate on whether this list of essential facilities (or ultimately essential services) should be universal or should vary based on characteristics such as age, gender, or other characteristics. Thus, the issue is not only whether destinations are accessible but also whether they should be for all citizens. This question (further explored in Section [2.1.2\)](#page-25-0) is not yet a settled matter, but has been paid attention in literature. Particularly, the study conducted by Radics et al., [2024](#page-114-0) presents a thorough review of articles analysing the topic and proposes a list for these services. In addition, the different transport networks in rural and urban areas also suggest that it may be required to develop different methodologies for each kind of area. Adding to that, a main requirement to align with the European directives is to aim for sustainability. Therefore, research should aimed to improve and promote the more sustainable alternatives of transport available for each case (see Section [2.3.2\)](#page-39-0). Considering accessibility, this would mainly translate in studying active mobility in the case of urban areas, and mainly public transport (or alternatively sustainable private vehicles) for the case of rural areas. In the latter case, public transport should be prioritized as "life cycle-based studies endorse public transport to cause lower environmental pressures compared to a private car" (Sinha et al., [2019,](#page-114-5) p.1), and recent studies further show that the averaged trip emissions are lower in public transport (Ritchie, [2023\)](#page-114-3). Finally, the last big ambition of this project is to establish indicators that are not only insightful for decision-makers but also actionable. Without proper assessment tools, it becomes difficult to identify those most affected and implement targeted solutions. These indicators should facilitate informed and data-driven decision-making and enable policymakers to anticipate the consequences of their actions. By creating effective metrics, it will be possible to implement targeted interventions that address transport poverty efficiently and equitably. # 1.3 Research question and sub-questions Considering all the above mentioned factors, the main research question is: How can the accessibility dimension of transport poverty be measured and presented, so accessibility measurements become useful for decision-makers when allocating transport investments? There are also several sub-questions that need to be answered. - How does (the lack of) accessibility contribute to transport poverty? - How is accessibility commonly measured? Is the current set of indicators suitable to assess it in the context of transport poverty, or are new indicators needed? What data is it needed to quantify accessibility in this context? - Is it possible to offer insights on the accessibility and transport poverty levels of rural areas? # 1.4 Academic and societal relevance # 1.4.1 Academic relevance From an academic perspective, this study aims to contribute significantly to the field by: • Refining the definition of transport poverty and the accessibility dimension. This research examines various definitions of transport poverty found in the literature and proposes a comprehensive working definition, aligned with the one provided by the EU commission. Additionally, it enhances the well-established concept of accessibility by contextualizing it within transport poverty studies. - Presenting a sequential methodology for analysing transport poverty. This work introduces a structured and sequential approach to analyse the different dimensions of transport poverty, enabling a clear distinction between the specific challenges associated with each dimension. . - Proposing a new accessibility indicator suited for the context of transport poverty. Drawing upon existing literature and original insights, this research suggests an indicator to assess the accessibility dimension of transport poverty. The indicator is explored in detail, demonstrating its applicability and relevance, and is also compared to existing ones. # 1.4.2 Societal relevance This study aspires to have a meaningful impact on society by: - Addressing transportation challenges to prevent transport poverty. Principle 20 of the European Pillar of Social establishes that everyone has the right to access essential services. In addition, adequate transport infrastructure is a precondition for EU competitiveness. Research also shows that transport plays a vital role in promoting economic development. It can contribute to enhancing intergenerational fairness and solidarity. (European Commission, [2025\)](#page-111-2). Therefore, efforts should be invested in preventing transport poverty from happening, and for that, it is first required to find ways to measure and identify transport poverty. This thesis works in that direction. - Ensuring that no on is left behind: This research emphasizes the need to ensure that rural communities (identified as particularly vulnerable when talking about transport poverty) are not left behind in the transition towards sustainable mobility. This work seeks to determine which areas have insufficient accessibility levels, highlighting transport as a fundamental necessity, rather than an advantage or a desirable goal. In addition, the choices made in Chapter 4, present how the proposed methodology is suited to put in a foremost place the rural areas. - Providing a framework to define, measure, and discuss transport poverty. A central objective of this research is to break down the abstract and complex concept of transport poverty into smaller, more tangible components, making it more accessible and comprehensible to the general public. - Offering quantitative indicators to help understand the severity of accessibility issues in the context of transport poverty. This study aims to equip decision-makers with analytical tools, particularly concerning accessibility, that enable them to make informed choices, thereby facilitating the prioritization and allocation of resources to cities most in need. Better information should lead to better decisions and therefore, eventually, result in a better society. - Use and merge different sources of information into a single, understandable indicator In recent years, more and more information is available. However, it is not always clear why this data is needed and in some cases, it has no use at all. This study aims to showcase that data collection is key to make better and more informed decisions and to prove how data from different sources can be merged and used to provide useful insights. # 1.5 Research design In the first Chapter of the work, an introduction to the problematic of transport poverty and in particular its accessibility dimension is presented. Then, the research questions and subquestions are exposed, as well as the academic and societal relevance of the study, expressing that it is important to analyse this project from a sustainable perspective and paying attention to vulnerable groups. Chapter 2 explores the literature related to transport poverty. This includes definitions about the different concepts of transport poverty, accessibility and rural areas, and a review of how this phenomenon has been measured in other publications and what tools could be of use to quantify this phenomena. It then briefly exposes the most important regulations related to the topic. Chapter 3 clarifies the methodology used. In this Chapter, the concept of accessibility is linked and placed in the context of transport poverty, and a comprehensive indicator (building upon Radics et al. [\(2024\)](#page-114-0) proposal) to measure accessibility is proposed, explaining how to use it in detail. Finally, a short explanation about the code used for computing the indicator is provided. Having a clear methodology, Chapter 4 presents a stud-case in Tarragona, to operationalize the described methodology. The objective of this example is to showcase the kind of results that could be obtained by applying the methodology, rather than to offer actionable results, as some simplification were made to reduce computing time. Chapter 5 unravels a discussion, investigating the strengths and limitations of the proposed indicator, how could it be improved or modified to offer a more comprehensive vision, and detailing how could it be replicated in other countries or regions (taking as an example case the Netherlands), and discussing what other results it could yields. Finally, this chapter concludes the work by summarizing the main insights and suggesting directions for future research. The research design has been summarized in Figure [1.1.](#page-21-0) ![](_page_21_Figure_1.jpeg) Figure 1.1: Research design of the thesis. # Chapter 2 # Literature review Transport poverty is an evolving concept that encompasses a wide range of diverse issues. Consequently, this chapter begins by examining the various definitions of transport poverty that can be found in the literature, along with the definitions of its principal dimensions, with a particular focus on accessibility. An additional short discussion on rural areas definition is provided, because of the particular relevance transport poverty has in these areas. Then, the chapter reviews how previous publications have measured and quantified accessibility. Given that accessibility is frequently associated with the analysis of travel time between origins and destinations, a concise section is then dedicated to explaining the Urban Transport Planning (UTP) model, as it could prove useful to compute them. Finally, the chapter briefly discusses some particularities related to transport decisions, and presents an overview of the main relevant regulations related to transport poverty and their primary implications. # 2.1 Definitions ### 2.1.1 Transport poverty Transport poverty has gained increasing relevance in recent years. However, the wide range of problems it integrates have led to various definitions being proposed in the literature. Therefore, several publications defining transport poverty have been identified, and are summarized in Table [2.1.](#page-23-0) Considering these definitions, it can be concluded that the concept of transport poverty is associated with a lack of: - Availability of means of transportation. - Transportation options that provide access to essential facilities within a reasonable (not excessive) amount of time. - Affordable transportation options that can be used to meet all mobility needs. - An adequate transport system that is safe, provides information, and is accessible to passengers with reduced mobility. - A mobility system aligned with the assumption of high mobility on which modern society is based. This is a system that considers spatial context, the labour market, and housing distribution. Table 2.1: Definitions of transport poverty from various sources. | Definition | Key words | Source | |--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------| | An individual or household is in transport poverty when
they do not have (suitable) public or private transport
(options) available to them and/or when the transport
system limits access to (other) essential goods and services
and/or when they have difficulty or are unable to meet
the costs of transport. | Availability,
accessibility,
affordability | Cludius
et al.,
2024 | | Individuals' and households' inability or difficulty to meet
the costs of private or public transport, or their lack of or
limited access to transport needed for their access to
essential socioeconomic services and activities, taking into
account the national and spatial context. | Affordability,
accessibility,
spatial context | European
Parliament
and
Council,
2024 | | "Limited transit options, compounded with socioeconomic
disadvantage preventing travel to important destinations,
like employment opportunities." (p.1) | Availability,
affordability, key
destinations | Allen and
Farber,
2019 | | The result of the direct and indirect interaction of
transport disadvantage (Fear of crime, no information,
high cost fares, poor services, no car) and social
disadvantage (low income, no job, no skills, illness, poor
housing). It is possible to be socially excluded but still
have good access to transport or to be transport
disadvantaged but highly socially included. Inaccessibility
is seen as a consequence of transport poverty, and not as a
cause. | Adequacy,
affordability,
availability, social
dimension | Lucas,
2012 | | "An individual is transport poor if, in order to satisfy
their daily basic activity needs, at least one of the
following conditions apply. There is no transport option
available that is suited to the individual's physical
condition and capabilities. The existing transport options
do not reach destinations where the individual can fulfil
his/her daily activity needs, in order to maintain a
reasonable quality of life. The necessary weekly amount
spent on transport leaves the household with a residual
income below the official poverty line. The individual
needs to spend an excessive amount of time travelling,
leading to time poverty or social isolation. The prevailing
travel conditions are dangerous, unsafe or unhealthy for
the individual." (p.1) | Availability,
individual
physical
condition, fulfil
daily needs,
affordability,
excessive travel
time, safety and
health | Lucas,
2018 | | Mobility-related exclusion (equivalent to transport
poverty) is "the process by which people are prevented
from participating in the economic, political and social life
of the community because of reduced accessibility to
opportunities, services and social networks, due in whole
or in part to insufficient mobility in a society and
environment built around the assumption of high
mobility" (p.211). | Accessibility,
assumption of
high mobility | Kenyon
et al.,
2002 | | Definition | Key words | Source | |--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|----------------------------------|--------------------------------------------------------------------| | Other components are also relevant when considering
transport poverty. For instance, housing costs (in the US,
cities with more commuter mode diversity are associated
with higher home property values). Labour markets can
also interact with transport poverty. | Labour market,
housing market | Frederick
and
Gilderbloom,
2018, Crisp
et al.,
2017 | Based on the aforementioned literature, it can be seen that transport poverty is a multifaceted issue encompassing problems of various natures. Consequently, it is common practice to compartmentalize the issue into distinct parts, leading to the identification of four dimensions of transport poverty as provided by the report of Cludius et al. [\(2024\)](#page-109-0) demanded by the European commission, illustrated in Figure [2.1.](#page-25-1) Each dimension pretends to concentrate a different kind of issue, allowing to group similar problems and differentiate them from others also related to transport poverty. The four dimensions are availability, accessibility, affordability, and adequacy. This categorization facilitates the disentanglement and separate study of the problem, thereby enhancing the understanding of transport poverty. The four dimensions are: - Availability: this dimension focuses on determining which means of transport are available for the population. For instance, it investigates if citizens own a private vehicle or have a public transport station within a reachable walking/cycling distance. - Accessibility: This dimension investigates if the means of transport provide access to essential services or goods within a reasonable amount of time. - Affordability: This dimension relates to financial issues, and investigates if transport options are inexpensive enough (in relation to income) to cover transport costs. - Adequacy: This dimension is transversal to the others, and mainly focuses on considering if transport options are safe enough, if they provide users with sufficient information on how to use them, and if they are accessible to passengers with reduced mobility. While this separation is relatively accepted, it remains an unresolved issue to determine when these dimensions are adequately satisfied. Specific requirements or indicators must be established to ascertain when each dimension is sufficient. For instance, a requirement for determining the sufficiency of affordability is proposed by Carruthers et al. [\(2005\)](#page-109-5): affordable transport is defined as transport that can be used at a price that is reasonable relative to income and does not undermine the ability to engage in other important activities. This could be operationalized for example establishing that transport associated costs should not be over 10% of the total income. Similarly, requirements for each dimension need to be clearly defined to obtain a complete understanding of transport poverty. When all these requirements are met, it can be concluded that all dimensions are fulfilled, and there is no transport poverty. Therefore, the following working definition is proposed: Transport poverty is the result of lacking availability, accessibility, affordability or adequacy, or a combination of them. This point of view incorporates all key elements found in previous literature, except for the explicit inclusion of labour and housing markets, as their consideration would overly complicate the framework. ![](_page_25_Figure_1.jpeg) Figure 2.1: Transport poverty dimensions, based on Cludius et al. [\(2024\)](#page-109-0). As a final notion, the current definition of transport poverty, does not emphasize directly the poverty dimension of its name, but this concept is present throughout all definitions. In Europe, poverty is conceived according to this cite: "People are said to be living in poverty if their income and resources are so inadequate as to preclude them from having a standard of living considered acceptable in the society in which they live" (European Parliament and Council, [2003,](#page-111-11) p.7). In this context, transport could be understood as a resource, that should provide enough access to services so that people can live in "acceptable conditions in the society they live". This underlying idea of enough can be found throughout all the definitions. For instance, when talking about affordability, the notion of "high expenditures in relation to income" highlights that transport should be cheap enough to live in acceptable conditions. Similarly, in the case of adequacy, transport alternatives should be safe, informative and adapted enough to be usable by citizens. Focusing in accessibility, this poverty dimension is included by only referring to the access to "essential services". By including this phrasing, it can be understood that not all trips to all destinations will be part of the assessment, but only those providing access to essential services, that all individuals should have access to. This is because "access to essential services is key to a full participation in society (...)" (Cludius et al., [2024,](#page-109-0) p.11). Based on this statement, essential services can be defined as those that need to be accessed by an individual to fully take part in society. This concept is tightly related to poverty, as not having access to essential services will preclude individuals from having a standard of living considered acceptable in the society in which they live, and therefore provoke poverty, in this case, transport poverty. To sum up, it is important to remember at all stages that the focus is on transport poverty, and therefore the that the poverty concept should be borne in mind throughout the whole work. ### 2.1.2 Accessibility in the context of transport poverty Given the significance of accessibility in the present study, a review of the literature was conducted to identify key contributions that provide accessibility definition. Four notable works were identified, each offering valuable insights into the concept. First, Hansen [\(1959\)](#page-112-6) defines accessibility as the ease of reaching destinations. Expanding on this, Kong et al. [\(2021\)](#page-113-6) conceptualize accessibility as the relative ease with which individuals can reach their desired activities. Third, Morency et al. [\(2011\)](#page-114-1) introduce the idea that accessibility is inherently linked with travel, and that for those who do not travel, accessibility is not defined. In their work, they also add the notion that, for those individual travelling, accessibility is related with the spatial dimension and distribution of services. Similarly, the fourth study Guagliardo [\(2004\)](#page-112-0) defines spatial accessibility as a measure of the ease with which individuals can travel from an origin to a destination using specific transport modes or networks. From the existing literature, it can be concluded that accessibility refers to the ease of reaching specific locations, is inherently linked to travel, and is a consequence of the spatial distribution of an area. However, considering these definitions, it is difficult to quantify accessibility. This is mainly a consequence of the fact that there is not a clear guidance or agreement on which locations should be analysed, a matter that remains unsettled. For instance, in Allen and Farber [\(2019\)](#page-109-4), access to workplaces is used as a proxy for overall accessibility, placing particular emphasis on employment-related destinations. Yet, in other studies such as Delmelle and Casas [\(2012\)](#page-110-3) and Sustrans [\(2016\)](#page-115-2), accessibility is assessed based on access to other facilities such as hospitals, libraries, recreational facilities, general practitioners clinics, post offices, or retail centres. In these studies, facilities selection is primarily driven by data availability and the selected facilities are assumed to provide a reasonable, if imperfect, representation of accessibility levels. While the specific choice of destinations may not significantly impact a general accessibility analysis—where the goal is to identify, in general, areas with higher or lower access to services—it becomes a crucial issue in the context of transport poverty. This distinction arises from the fundamental objective of transport poverty analysis: to evaluate if the the accessibility levels are enough, this means, if they prevent individuals from fully participating in society (European Commission, [2017,](#page-110-4) Cludius et al., [2024,](#page-109-0) Council of thr European Union, [2003\)](#page-110-5). This is a result on its own, and does not necessarily need comparison with other regions. Identifying essential destinations therefore becomes of paramount importance. For instance, while libraries or post offices may serve as useful proxies for accessibility in general— as they are often located in city centres or near commercial hubs—they may not be appropriate indicators for transport poverty. In this context, lacking access to a hospital within an hour could reasonably be considered a form of deprivation, whereas the same argument would be difficult to justify for a library. This distinction highlights why traditional accessibility indicators may fail to adequately address accessibility challenges in the context of transport poverty. Building on these perspectives, it is possible to revisit the proposed definition in Section [2.1.1:](#page-22-2) accessibility refers to the ability to reach essential services (provided in essential facilities) within a reasonable amount of time. This definition integrates key elements from the literature. First and foremost, the inclusion of only the essential services contextualizes in the context of trasnport poverty. It encompasses Hansen [\(1959\)](#page-112-6) concept of ease of access, while also refining the notion of desired activities to essential activities to align with the study's focus on transport poverty. Furthermore, the inclusion of a reasonable amount of time frame implicitly accounts for the origin-destination framework outlined by Guagliardo [\(2004\)](#page-112-0). Yet, the working definition still has two elements that need to be established. The list of essential facilities (providers of essential services), as discussed above, and of a reasonable amount of time. Both elements are further discussed in the methodology Chapter [3.](#page-44-0) ## 2.1.3 Rural areas Given the particular attention paid to rural areas in this study, it is important to understand how these areas are defined. In Spain, the operative definition is established by national legislation. According to the Cortes Generales de Espa˜na [\(2007\)](#page-110-6), a municipality is considered rural if it has less than 30,000 inhabitants and a population density below 100 inhabitants per square kilometre. This definition is straightforward and commonly applied in national rural development policies. However, rural classifications vary across institutions and countries. Eurostat, for instance, uses the Degree of Urbanisation (DEGURBA) to define rural areas based on population density and settlement structure rather than administrative boundaries. Under this system, a territory is classified as rural when more than 50% of its population lives in grid cells identified as rural—typically with fewer than 300 inhabitants per km2 or fewer than 5,000 total inhabitants (Eurostat, [2021\)](#page-111-12). These cells must also lie outside of larger urban clusters. While this method facilitates cross-country comparison within the EU, it can be more complex to implement at the local level and may be less familiar or accepted by national policymakers who rely on domestic administrative classifications. Beyond the Spanish and Eurostat definitions, many other systems exist. For example, Cattaneo et al. [\(2021\)](#page-109-3) describes several classification schemes used in the United States, such as the Rural-Urban Continuum Codes (RUCCs) developed by the USDA, the Urban Influence Codes (UIC), and the Rural-Urban Commuting Area Codes (RUCAs). Similar variations can be found in other countries, each emphasizing different criteria. This diversity of approaches highlights that rural-urban classification is not a universally fixed concept but rather one that depends on the context and objectives of the analysis. # 2.2 Measuring accessibility in transport As seen in previous Section, accessibility is one of the four main transport poverty dimensions, and is related to the capability of accessing essential facilities in a reasonable amount of time. In this section, attention is paid into investigating how the existing literature has so far analysed the topic. To that end, several publications are examined, exposing the methodology used and discussing its strengths and weaknesses. # 2.2.1 Currently used indicators A comprehensive approach to measuring accessibility must consider both individual and systemic factors. On the one hand, the systemic part holds great influence, as the network design, the available infrastructure and the transport planning highly influence travel times and accessibility. Therefore, numerous publications propose indicators related to it. However, accessibility is also influenced by personal preferences, mobility constraints and socio-economic conditions. In addition, societal factors such as transport systems, crime rates, economic conditions, and regulatory frameworks further shape accessibility (Kong et al., [2021\)](#page-113-6). These factors are crucial in the context of transport poverty, where limited mobility restricts access to essential services, reinforcing socio-economic inequalities. Therefore, a comprehensive measurement should integrate both personal and systemic dimensions, capturing personal conditions and systemic factors. However, when analysing the methodologies currently applied in accessibility research, a discrepancy emerges between these theoretical recommendations and practical implementations. Notley, personal preferences and conditions are not often considered, and only the systemic part of the analysis is analysed. To better understand how accessibility is measured in practice, this section examines eight key studies (which have been classified into three groups are categories) that explore accessibility from different perspectives. Some of these studies focus explicitly on analysing geographic accessibility, while others use alternative measurements as proxies. By reviewing these approaches, this analysis aims to identify the most relevant methods and indicators for the context of this research. ### 2.2.1.1 Accessibility based on travel time to key destinations In these studies, accessibility is mainly measured by reflecting how long it takes citizens to reach a set of key destinations. Studies in this sense are usually based on GIS data and involve geo-calculations (or an equivalent measurement) to find out how long would it take to reach key destinations using several ways of transport. Four case studies are synthetized below. EU Commission report, 2024 In the EU Commission report Cludius et al. [\(2024\)](#page-109-0), only one indicator is calculated to measure accessibility: One-way commute to work of more than 30 minutes. This indicator focuses in only one essential destination (in this case the workplace), and suggest a reasonable time of 30 minutes. With the available data (Labour Force Survey 2019, module on work organization and working time arrangements) the authors are capable of defining the share of active population spending more than 30 minutes commuting to work. This measure has the advantage that it is computable for Spain, but is not very actionable, as it is difficult for decision makers to apply concrete measures to decrease this number. Canada 2019 measurement of accessibility In a study conducted in Canada (Allen and Farber, [2019\)](#page-109-4), the authors posit that job accessibility can be used as a relevant indicator for measuring accessibility levels: "Job accessibility is a good proxy for access to other destinations (...) like shops, services, and recreation"(p.3). Therefore a measure of job accessibility is proposed and developed. Unlike hospitals, which centralize health services in specific locations, workplaces are dispersed across various sites. This dispersion complicates the task of identifying the appropriate destination for analysis for each citizen, a step needed to compute the average travel time and thus accessibility to work. To address this issue, the authors propose using a gravity model to achieve a realistic distribution of job destinations. By implementing the model they propose, the authors can create an origin-destination (O-D) matrix, which establishes the number of people travelling from each area of study to various work locations. The model is then enlarged to account for different travel modes (in this case car and public transport), resulting in equations (1), (2) and (3) of the paper of Allen and Farber [\(2019\)](#page-109-4). The proposed formulation, that is computed with an iterative algorithm, allows for the assessment of job accessibility by incorporating the fact that not all workers work at the same place, but that they compete for job locations. The authors also provide a publicly accessible GitHub repository with the code used to compute it, free to use. An important advantage of this indicator is that it only requires aggregated data from regions. Another one is that it takes into account the competition for jobs present in other surrounding living areas, which is realistic. Additionally, the availability of the code used to compile is also an asset. A downside is that it is difficult to predict what will be the result if a given parameter is changed (since it is the result of an iterative process) and that the model needs calibration, requiring data that is not always available. Considering all this, it seems a promising indicator to be used. As a curiosity, they find that in Canada, lower income neighbourhoods tend to have better levels of accessibility. Accessibility case study in Cali, Colombia 2012 In the case study conducted by Delmelle and Casas [\(2012\)](#page-110-3), the accessibility improvements resulting from the implementation of a new LRT (Light Rail Transit) bus line in Cali are analysed. The objective of the LRT proposal is to ensure that all residents experience an equitable level of accessibility upon its implementation. The authors therefore need to deign an accessibility measure to check if it is the case. Two key factors are highlighted in their analysis. First, citizens must be able to reach the LRT bus stops by foot. Although the study refers to this aspect as accessibility, following the conceptual framework outlined in Figure [2.1,](#page-25-1) this would be classified under the availability dimension. Second, individuals must be able to reach essential destinations efficiently. In the study, the destinations analysed are hospitals, libraries and recreation facilities. This aligns with the accessibility dimension in Figure [2.1.](#page-25-1) To measure accessibility, the study employs Hansen's (1959) gravity-based model, given by Equation [2.1:](#page-29-0) $$A_i = \sum_{j=1}^{n} S_j d_{ij}^{-\beta} \tag{2.1}$$ where Ai represents the accessibility index for an individual i, Sj denotes the attractiveness of destination j, dij corresponds to the distance or travel time between locations i and j, β is a distance-decay parameter controlling the influence of distance, and n is the total number of destinations. The study also integrates a temporal dimension to account for variations in demand over time. Because of the complexities inherently linked to analysing public transport, the authors apply simplification. For instance, one simplification made in their approach is that individuals always choose the nearest station. Further simplifying the model, waiting and transfer times are not explicitly considered in the model. Another assumption under this formulation is that the trip generation depends solely on transport supply, without accounting for demand. Alternative models introduce variations that address these limitations. A significant advantage of this approach is that it does not require origin-destination (O-D) matrices, although they can be used for calibration. Additionally, it does not rely on individual-level data, as accessibility is calculated from centroids to specific destinations. Furthermore, the model considers all potential destinations in the final accessibility measurement, ensuring that even distant locations contribute to increasing accessibility—albeit with lower weight—to the total accessibility score. However, there are important limitations. The results from accessibility measures derived from this study are not directly comparable to those from other studies, as the chosen parameters can vary. Additionally, the model does not account for work accessibility, which is a crucial essential destination according to the two previously exposed publications. Furthermore, it does not incorporate population density, meaning that the number of residents affected by poor accessibility is not considered. This limitation is particularly relevant for prioritization: an area with extremely low accessibility may be of lower priority if it has a very small population, whereas a moderately inaccessible area with a high population density might require more urgent intervention. Thus, including population data and travel frequency could be helpful to improve the accessibility assessment. Case study in Scotland, 2016 In the report by Sustrans [\(2016\)](#page-115-2), an indicator for the risk of transport poverty is developed. The authors define transport poverty risk as occurring in areas with high travel times (more than 1 hour) and low availability (less than 2 frequencies per hour). The study identifies that areas at high risk of transport poverty are characterized by low income, high car ownership, and limited access to essential facilities. The inclusion of high car ownership as a risk factor may seem counter-intuitive; however, it is based on the assumption that elevated car ownership rates result from inadequate public transport services, thereby compelling individuals to own vehicles and bear excessive costs, ultimately leading to transport-related financial strain. Once again, accessibility is based on the travel times to access key destinations. In this case, the study utilizes the SIMD dataset, which provides average travel times to key facilities by both private and public transport. The key facilities considered in the study include: General Practitioner (GP) services, Post Offices and Retail Centres. Regarding accessibility measurement, the report classifies accessibility levels using a threetiered scale: - Low (Score 1): All three essential facilities are reachable within one hour. - Medium (Score 2): At least one of the facilities requires more than one hour of travel time. - High (Score 3): Two or more essential facilities require more than one hour of travel time. Additionally, the study incorporates an indicator for public transport availability, classified as follows: - Low (Score 1): Large urban areas and urbanized regions. - Medium (Score 2): Small towns in remote areas with 1–2 services per hour or small accessible towns with 1–4 services per hour. - High (Score 3): Rural areas with only 1–2 services per hour. To determine overall accessibility poverty, the study sums both measures and classifies risks (sum = 2:low, sum = 3 or 4:medium and sum=5 or 6:high). Using these indicators, the authors produce a colour-coded map identifying areas at risk of transport poverty. This methodology presents several advantages. It is straightforward, easy to interpret, and computationally efficient, making it accessible for both researchers and policymakers. Additionally, it provides a systematic way to rank cities based on transport poverty risk, allowing authorities to prioritize interventions effectively. Despite this, the approach has several limitations. It does not account for job accessibility, which is arguably the most critical determinant of transport equity (job places are not explicitly included in the SIMD dataset). Furthermore, the criteria used to assess public transport frequency are subjective. For instance rural areas are broadly classified as high-risk without considering whether some of them might be well connected to urban centres. Another limitation lies in the constraints of the dataset, as it only allows for computations based on predefined destinations, restricting the ability to evaluate accessibility to other locations. Lastly, the classification thresholds, such as the one-hour reference time, are not well justified or explained, raising concerns about their validity. ### 2.2.1.2 Accessibility as a mix of transport related indicators In this case, accessibility is not straightforwardly analysed, but rather inferred from the analysis of mobility related indicators. Two publications in this direction are detailed. Case study in three cities in Canada, 2011 In their study, Morency et al. [\(2011\)](#page-114-1) make use of a highly detailed dataset containing precise individual-level information from 5% of all households, in order to predict the distance travelled by different age groups. This distance serves as a proxy for accessibility, as people with higher travelled distances will be able to reach more services and therefore have higher levels of accessibility. For each individual, the dataset records the exact locations they visited, represented by x-y coordinates, as well as the specific transportation modes used, including subway stations and road routes for car travel. This dataset is derived from two of the largest cross-sectional surveys in the world, providing a robust foundation for their analysis. The primary methodological approach employed in the study is multivariate regression. Given the availability of explicit spatial coordinates, the authors extend their analysis by developing a spatially-expanded regression model. This model enables them to capture approximately 20% of the observed variability and to examine differences across various population groups. This case study underscores the depth of analysis that can be achieved when comprehensive datasets are available. Additionally, it serves as a valuable reference for designing an interview protocol capable of collecting sufficient information for transport accessibility research. A key advantage of this study is that travel times and destinations are directly obtained from the dataset rather than estimated. As a result, the use of Geographic Information System (GIS) tools becomes unnecessary, simplifying the analytical process. Yet, the replicability of this study is very low, as the dataset used is not commonly available. Another disadvantage of this approach is that it is not possible to infer which areas are more in need. Spatial disparity of transport in Cali, Colombia, 2012 The study by Jaramillo et al. [\(2012\)](#page-112-7) is notable, although not directly related to accessibility, due to its use of a variety of indicators that are combined with different weights to create a comprehensive, singular indicator. The primary objective of this paper is to develop two indicators: one representing the need for public transport and the other indicating the provision of public transport. By comparing these two indicators, the study aims to identify areas that require more attention in terms of transport provision. The transport needs indicator is calculated as a weighted sum of factors such as lack of private vehicle ownership, old age, disabilities, unemployment, and access to key destinations like the city centre, educational services, and the degree of insecurity (see Table 1 in Jaramillo et al. [\(2012\)](#page-112-7)). The transport provision indicator is calculated by considering the number of transport stops times their capacity times their frequency, divided by the area of the region (and population in the case of the relative measure). This results in a unit of flow: [P eople/h ∗ m2 ]. Both indicators are later on normalized (so they range from 0 to 1), eliminating the units and making them dimensionless. These indicators are then compared to assess the relative levels of need and provision. Given the large number of indicators considered in the need for transport calculation, not all indicators contribute equally to the overall model. To identify which indicators are most relevant, the study applies Principal Component Analysis (PCA), a widely used technique for data reduction. PCA analyses the data and identifies which variables account for the most variance. The principal components are linear combinations of the input variables, and by selecting only the components that significantly influence the model (usually PC1 and PC2), it is possible to reduce the number of variables and assign different weights based on their relevance. Through this technique, the authors reduce the initial set of 15 indicators to 9, each weighted according to its significance. One of the main advantage of this formulation is that the results are intuitive. For example, areas with high unemployment, an ageing population, and a significant distance from the city centre are awarded high need for public transport. In contrast, areas with low bus frequency and capacity will show a low level of public transport provision. Additionally, it allows to properly rank and see which areas are more needed in public transport, or which areas are better or worse provided by public transport. Despite the strengths of this methodology, several limitations exist. The use of normalized indicators is a key issue, as it does not allow for an assessment of whether public transport provision is actually adequate. While the method identifies areas with high or low need and provision, it fails to determine whether the available transport meets the actual demand. For instance, it is conceivable that a country or city could have an overall shortage of public transport. In such a case, even areas with a relatively high level of provision compared to others (near a score of 1) could still be insufficient to meet the needs of the population, even in areas with lower needs. Furthermore, while the methodology is effective at identifying which areas require the most urgent attention, it does not specify the types of interventions needed. For example, while it can pinpoint areas where more buses are needed, it cannot determine exactly where these buses should be headed towards. Although not directly related to any study, other similar indicators related to accessibility that would fall in this category are: - Number of public transport services leaving the city per hour - Number of destinations reachable by public transport - Among the five closest cities (with over 10,000 inhabitants) to a rural area, how many can be reached without transfers by public transport - Commuting time to the nearest city with more than 500,000 inhabitants - Number of hours public transport services are available ### 2.2.1.3 Accessibility based on access to essential facilities Finally, a third type of analysis is presented. Here, accessibility is defined in terms of access to essential destinations—a shift from purely mobility-based analyses to a needs-oriented approach. This view aligns greatly with the proposed definition, focused on accessibility to essential facilities in a reasonable amount of time. Two particularly influential works in this domain have been identified. Accessibility to essential facilities in Sevilla, Spain This work of Radics et al. [\(2024\)](#page-114-0) presents a detailed analysis of daily accessibility to essential facilities in Sevilla, a city in Spain. The study focuses on two means of transport, walking and cycling, as the study is conducted inside a city. One of the main issues faced by this case study (apart from the geospatial calculations) is the establishment of a list of essential facilities that should be considered in the accessibility analysis. To overcome it, authors ground their choices in a comprehensive literature review aimed at identifying which destination categories are necessary to ensure social inclusion and daily functioning. Based on this review, six key categories of essential facilities are defined: - Transport stops: in this study transport stops are treated as essential destinations people need access to. It is noted however that, according to the definitions section, access to transport would fall under the scope of availability analysis, not accessibility. - Working locations: this category is included in the recommended classification, but a comment is said that because of the difficulty of gathering this data, "working is not considered in most of the reviewed papers". It groups locations that act as work trip attractors, such as Central Business Districts (CBD), or the headquarters of important companies. - Commerce: This category includes food stores, fresh-food shops, and markets, recognizing that access to nutritious and affordable food is an essential necessity. - Healthcare: health facilities encompass healthcare centres, hospitals, pharmacies, and social care facilities, reflecting the importance of regular and emergency medical care. - Education: this category of essential facilities covers all institutions where all stages of education are included, emphasizing the right to education and the role of schools in social cohesion. - Leisure Facilities: This category includes parks, playgrounds, sports facilities, libraries, and civic centres, acknowledging that social participation and well-being are essential components of a liveable environment. This classification offers a robust and operationalize basis for defining essential facilities in accessibility studies. Apart from that, this publications also offers an interesting point of view for analysing accessibility. The authors highlight the significance of setting time thresholds for accessing essential facilities. In Table 2 of Radics et al. [\(2024\)](#page-114-0), specific time thresholds are defined for each destination. Consequently, accessibility assessment is conceived as a process of measuring the travel time required to reach each essential facility and of verifying if this time falls within the acceptable time threshold, as presented in Table 3 of their study. Considering these points the authors are capable of outputting relevant figures and maps, highlighting which areas have better or worse accessibility to facilities in the city of Sevilla, and providing highly informative outputs of the regions more in need. A downside of this publication is that it only considers active mobility (walking or cycling), and not private or public transport, and that personal preferences or differences across groups are not captured. Transport equity, accessibility to essential facilities for more resilient cities The second reference is Logan and Guikema [\(2020\)](#page-113-7), which, although situated in the context of resilience, reframes accessibility as a core component of equitable cities, including the notion of equity. The author argues that resilience is inseparable from accessibility, especially in scenarios of disruption such as natural disasters. In this sense, accessibility is not only about present-day provision but also about ensuring that essential facilities remain reachable under adverse conditions. This term of transport equity introduces the view that the impacts (benefits and costs) derived from transportation should be distributed fairly, and that "planning decisions can have significant equity impacts" (Litman, [2017,](#page-113-8) p.1). Other publications in the subject also conclude that although "there is not yet an appraisal method that adequately reflects transport equity issues" it is important to account for equity. To sum up, the transport equity term emphasises how transport should enable different people to flourish and live meaningful lives (Pereira, Karner et al., [2021\)](#page-114-6). This makes it a matter tightly related to transport poverty. Coming back to the article of Logan and Guikema [\(2020\)](#page-113-7), the authors analyse the proximity of the population living in an area to essential destinations such as supermarkets, schools or service stations. Then, a simulation of a natural disaster is made, making some of the facilities to be closed, and leading to a new situation, with lower accessibility, as some of the facilities are closed. This provides a suitable framework to analyse how resilient cities are to disasters. Crucially, this work highlights the importance of contextual and community-based definitions of essential services, to establish which are the essential facilities. Rather than assuming a fixed list of universal needs, the author emphasizes the role of community engagement in identifying which services matter most in a given place. Nevertheless, the study points to a common set of core services stating that "accessibility of services such as education, healthcare, food access, and cultural amenities is crucial for a community's vitality, liveability, and cohesion" (p.1539). Together, these two works frame accessibility not merely as a matter of infrastructure or travel time, but as a social concept that seeks to guarantee equitable access to the services necessary for daily life. They provide both methodological guidance and moral justification for the indicator developed in this thesis. ### 2.2.2 Computing average travel time to decentralized services From the previous section, it is possible to infer that accessibility measurements are directly related to travel times (or distances) between household locations and the destinations under study. In general, when services are centralized—such as hospitals, universities, and shopping malls—a time or distance matrix is sufficient for conducting a proper accessibility analysis. To obtain it, it is needed to either have a source where these data has already been collected, or dispose of GIS tools (including coordinates of the places and network data) to compute them. However, in cases where services are not centralized -such as employment opportunitiesaccessibility analysis can become more complex. Multiple locations may offer the same service within a region, requiring not only an assessment of travel times to each location but also an understanding of how many individuals travel to each destination, to calculate an average commuting time. To address this, a brief introduction to the UTP model is proposed. It is noted that, while this section may offer good guidance for further developing the indicator, it has not been specifically used within the scope of this thesis. The UTP model Elements from the Urban Transportation Planning (UTP) model, originally developed in the 1960's, can be applied to calculate Origin-Destination (O-D) matrices. The UTP model, useful for transport planning and suited for traffic simulation, follows a four-step process: 1. Trip Generation Estimation 2. Spatial Distribution of Trips 3. Modal Choice (introduced later in the model's development) 4. Route Assignment (Willumsen et al., [2011,](#page-115-3)Zhou et al., [2009,](#page-115-4)Chu, [1990\)](#page-109-6) - 1. Trip Generation and Attraction: Estimating (Gi , Ai) To determine the number of trips generated (Gi) and attracted (Ai) by each area, various estimation methods can be employed. The primary goal of this section is to establish, for each area of study, the number of trips generated and attracted by it. A list of several methods that can be used it detailed, although more are available. - Growth Factor Method: This forecasting approach is useful when existing data on current trip generation is available. The method consist on escalating by a factor a previously known matrix, to fit the current number of total trips. Although easy to compute, this method is not able to incorporate mobility pattern changes, and requires a complete O-D matrix as a starting point. These method are "trivial, but mostly inadequate"(Willumsen et al., [2011\)](#page-115-3). - Category or Cross-Classification Method: This method is applicable if tables detailing the number of trips per household are accessible. By combining this information with the number of households of each type in each area of study, it is possible to get estimations for (Gi), and sometimes also for (Ai). This method is used, for instance in Stopher and McDonald [\(1983\)](#page-114-7), and can be used with current surveys such as the [EMEF](https://ce-sermetra.atm.cat/es/web/observatori/w/encuesta-emef) transport survey in Catalu˜na. - Regression Model: If data is available for certain rural areas (e.g., derived from road traffic measurements), a linear regression model can be developed, to infer the data about areas where no measurements are available. This model could additionally incorporate explanatory variables such as population size, car ownership rates, and the percentage of elderly residents, to estimate the number of trips generated and attracted by each area. - 2. Trip Distribution Modelling: Estimating (Tij ) Once the number of trips generated and attracted by each area is known, the next step is determining how they are distributed, this means how many people go from every origin areas to every destination area. (Tij ). Several modelling approaches can be applied, and in most of them it is required to know the cost (in terms of time or distance) of travelling from every origin to every destination. - Proportional Models and Singly/Doubly Constrained Models: If an OD matrix is available for a reference year, future distributions can be estimated by applying a proportional yearly adjustment factor. One technique to do so is with the Furness method (Furness, [1965\)](#page-111-13), which consists on performing row and columns adjustments, an algorithm that , except in special cases, converges quickly to a unique solution. • Synthetic Distribution Models: These models use mathematical functions to estimate trip distributions. A common formulation uses Equation [2.2:](#page-36-0) $$T_{ij} = \alpha_i O_i \beta_j D_j f(c_{ij}) \tag{2.2}$$ where Tij is the number of trips generated from i to j, Oi is the number of trips originated at destination i, Dj is the number of trips attracted by zone j, αi and βj are coefficients to ensure internal consistency, and f(cij ), known as the deterrence function, can have several forms (exponential, power, gamma, log-normal, etc.) and depends on the cost (usually distance or time) of going from i to j. The gravity model is one of the commonly used synthetic models for trip distribution. Inspired by Newton's law of gravity, these models estimate trip distribution based on the trip generation and attraction (equivalent to the mass), and distance between areas. A commonly used formulation is the Cobb-Douglas model: $$T_{ij} = kO_i^{\alpha} D_j^{\beta} c_{ij}^{-\gamma} \tag{2.3}$$ where k, α, β, γ are parameters calibrated to best fit the observed data. For more information, for instance the paper of Arasan et al. [\(1996\)](#page-109-7) investigates how synthetic gravity models work and can be calibrated, as well as other relevant books such as Willumsen et al. [\(2011\)](#page-115-3) or Meyer and Miller [\(2001\)](#page-114-8) • Entropy Maximization Models: These models aim to maximize entropy, providing a robust mathematical framework that enables sensitivity analysis and convergence proofs. An example of these formulations is the Wilson Model (Wilson, [2013\)](#page-115-5) , following Equation [2.2](#page-36-0) with a deterrence function defined as: $$f(c_{ij}) = e^{-\gamma \cdot c_{ij}}. (2.4)$$ Another instance is the model developed by De Grange (de Grange et al., [2012\)](#page-110-7), suitable for short trips, exposed in Equation [2.5:](#page-36-1) $$T_{ij} = \alpha_i O_i \beta_j D_j c_{ij} e^{-\gamma \cdot c_{ij}} \tag{2.5}$$ Parameter Calibration and Data Requirements Each of these models relies on parameters that must be carefully calibrated. Whenever possible, existing datasets should be utilized to that end, extracting as much as data as possible from the,. For calibration, the following data is required: (1) The number of trips generated and attracted by each area. (2) Travel costs between areas (typically in terms of time or distance) . Some specific approaches exist for this purpose. For instance, the : Hyman's Method ,requires only the average travel time of the population, typically obtained through surveys. Another possibility is to use the Triproportional Method, that enables to compute the friction factors and discover the best-fitting functional form using trip distribution data segmented by distance ranges (e.g., trips between 0–5 km, 5–10 km, etc.) (Codina, [2025\)](#page-110-8). By selecting the appropriate model and calibration approach, an accurate trip distribution matrix can be generated. This would allow, if needed, to obtain more precise measurements on average travel time to non-centralized essential services, which are not all provided in a same facility, such as work. 3. Modal Choice : The goal of this section of the UTP model is to identify the mode of transport each individual uses when travelling from their origin to their destination. After completing steps 1 and 2, the number of people travelling from each origin to each destination is known. However, there is no information on the mode of transport they use. This can be determined using Random Utility Models (RUM) (Daganzo, [2014\)](#page-110-9). Based on utility theory, RUM posits that individuals will choose the mode of transport that offers them the highest utility among all available options. The utility function is defined with a systematic part, that depends on observable factors that influence the decision making, such as the travel time of the trip for each alternative, its price, the income of the user, among others, and an error term, that captures the differences of perceptions of each individual that can not be contained in the systematic part. Depending on if this error term is assumed to be normally distributed or to follow a logistic distribution, a probit or logit model is used (National Cooperative Highway Research Program, [n.d.\)](#page-114-9). In both scenarios, using these models allows to predict the most likely mode of transport each individual will choose based on systematic variables, utilizing statistics and data. 4. Route assignment Finally, once it is known how many people will go from each origin to each destination and how, this fourth section focuses on analysing which route each individual will take. This fourth part is mainly based on computing shortest path routes. For computing these paths, one common assumption is to use the Wardrop's equilibrium. This concept assumes that users select a route that minimizes the time or cost incurred in its traversal. This behavioural assumption admits convenient mathematical descriptions, and efficient algorithms for the computation of equilibria are available. (Correa and Stier-Moses, [2011\)](#page-110-10). Other assumptions can also be applied. This shortest path algorithms is something broadly studies and very commonly used in software such as Google Maps(R) or AIMSUN (R), among others. At the end of these four steps, it is possible to gain a very broad understanding of travel behaviour, and based in data it is possible to compute how many people travel from each origin to each destination, their modal choice and the total travel time or cost it takes to do it. # 2.3 Decision making in transport Decision-making in mobility and transportation is a complex process with several unique challenges. Broadly, transport-related decisions can be categorized into two main types: high-level (strategic) decisions, which involve spatial planning, infrastructure investments, and transport system design, typically made by policymakers; and user-dependent decisions, which are influenced by individual preferences and behaviours.While both perspectives are essential for understanding transportation systems and are closely interconnected, this thesis primarily focuses on strategic decision-making. Therefore, in this section, an overview of the complexities faced by decision makers when adopting transport measures, and a short overview of the Spanish and European main regulations on the topic is presented. ### 2.3.1 Particularities of the decision making in transport A non exhaustive list of elements that need to be considered when analysing the public transport decision-making process is explained below. This list is aimed to understand why this topic is so complex and partially explain why it is not a trivial matter to solve it. - Public transport companies do not always to make profit. Public transport companies in Europe such as RENFE, SNCF, Nederlandse Spoorwegen (NS) among others, are public entities, and therefore do not (solely) aim to make a profit, but also to provide a public, social service. For instance, in 2023, RENFE, the Spanish railway company, received 1.5 M € in operating subsidies from the State General Administration (Renfe-Operadora (Grupo), [2023\)](#page-114-10). This makes decision indicators more complicated, as the objective is not to be lucrative or to always reduce cost. More elements need to be taken into account to decide if a policy is worth it. - Inclusiveness is a subject under debate. Inclusive policies often require from public funds. Therefore, two opposite mentalities collide. On the one hand, it is good to optimize the spending and minimize cost, so other social projects can also be funded. However, there is also the notion that enough money should be spent. One argument to support this is that exclusion generate inefficiencies that can be costly. For instance, in C´amara et al. [\(2020\)](#page-109-8), the authors analyse the problems associated to exclusion of disabled people from the labour force,considering that it results in high cost to society because of such exclusion. As a consequence, there is always an ongoing debate about the amount of money that should be spent in inclusive policies. - Transportation decisions are influenced not only by technical expertise but also by political considerations. While technical knowledge is taken into account, the chosen solution is not always the optimal or most efficient. This is due to the inherently political nature of decision-making, where public perception plays a significant role. As a result, while analyses must be technically sound and provide meaningful insights, they should also be communicated in a way that is accessible to policymakers and non-technical stakeholders. This is particularly important when designing indicators, as they need to be easily interpretable. For instance, an indicator stating, "This configuration will reduce travel time by 5%" is more effective than one that reads, "This configuration will increase accessibility by 5%." - Alignment with EU Policies and Spanish strategic planning. As a member of the European Union since January 1, 1986, Spain is required to align its national objectives and policies with EU directives, while also considering its own strategic planning. Consequently, indicators should be designed to assist decision-makers in evaluating whether their choices contribute to the region's progress in a manner consistent with both EU policies and Spain's national priorities. The Section [2.3.2](#page-39-0) explores forward on this topic. - Transport solutions are mainly associated with high investments. Every transportation decision must be carefully evaluated, as it typically involves a substantial initial investment and/or high operating costs. While some modes of transport are more expensive than others—such as trains or metros compared to buses—all transport-related projects generally require costly vehicles, continuous staffing, and extensive planning to design, implement, and launch the service effectively. While this is trying to be reduced introducing new mobility options, such as shared mobility or on demand transport options, in most cases and due to little or none previous experience, these projects are difficult to launch and require high promotional and education efforts before they can be implemented (for instance Burgerb¨us [in Germany,](https://ruralsharedmobility.eu/wp-content/uploads/2019/08/SMARTA-GP-BurgerBus.pdf) a local service for on demand transport or Vivolt [in France,](https://vilvolt.fr/) a shared bike solution for rural areas). - Solutions to accessibility may not be transport related. Another relevant aspect, that also adds complexity to the decision making, is that "solutions to transport poverty issues might not be transport related" (Metta, [2020\)](#page-113-3). For instance, accessibility to jobs can be improved by increasing home working, or accessibility to hospitals can be improved, either by connecting existing hospitals better, or by building new ones. ### 2.3.2 Main political strategies related to the topic As can be derived from previous section, transport is abroad and complex issue, and therefore it has been an important source of problems for society, that have derived in numerous policies. Consequently, several regulations from different strategic levels cover it. Below is a selection of the main policies at global, European, and Spanish level, addressing important concepts that need to be beard in mind when investigating transport poverty. ### 2.3.2.1 Global and European policies At global and European levels, two main types of documents are relevant for transport: general directives applying to all industries and fields, and specific transport-related regulations, providing targeted advice for designing transport policies. General Directives for all fields Four main elements have been identified as providing general guidance to any project that wants to be developed in the world and more specifically in the EU. These legislations mainly introduce the concept of sustainable development, expressing that all new research, studies and advances should allow to advance towards a more sustainable and fair society. - Sustainable Development Goals (SDGs): In 2015, all United Nations Member States adopted the 2030 Agenda for Sustainable Development (United Nations General Assembly, [2015\)](#page-115-6), providing general guidance for all projects (including research) from 2015 to 2030. The 17 SDGs are an urgent call for action by all countries in a global partnership to become more sustainable and reduce emissions. Several SDGs are strongly linked to transport poverty, including Goal 1 (No Poverty), Goal 10 (Reduce Inequalities), and Goal 11 (Sustainable Cities and Communities). - Green Deal and Fit for 55: In 2021, the European Commission [\(2021b\)](#page-110-11) was adopted, translating the European Green Deal (European Parliament and Council, [2019\)](#page-111-3) into law. The aim of the green deal is for Europe to become the first continent with no net emissions by 2050, a concept often referred to as the sustainable transition. To achieve this, the EU launched the "Fit for 55" package, a set of laws aiming to reduce greenhouse gas emissions by at least 55% by 2030. This package includes regulations as important as the EU Emissions Trading System (EU ETS), the Social Climate Fund (SCF), and the Carbon Border Adjustment Mechanism (CBAM), among others (European Council, [2021\)](#page-111-9). These regulations are crucial for all sectors, including transport, which must also become fully sustainable (net zero emissions) by 2050. In addition, these laws emphasise that this sustainable transition should be fair, leaving no one behind, and paying particular attention to vulnerable groups. - ETS 2 and Social Climate Fund (SCF): In addition to the EU ETS included in the "Fit for 55" package, which is the main regulatory element to incentivise the decrease of the Greenhouse emission, from 2027, a new, separate ETS will help reduce emissions from buildings, small industries, and road transport. This ETS 2 (European Parliament and Council, [2003\)](#page-111-11) aims to switch to low-emission transport, making it especially important for active mobility and public transport sectors. To support these changes, the Social Climate Fund (SCF) will mobilize €86.7 billion from 2026 to help vulnerable people transition towards a sustainable life style. This fund is crucial for transport, as it will finance projects that encourage sustainable mobility (European Commission, [2021a\)](#page-110-12). The SCF refgulation also includes defintions of transport poverty. - European Urban Agenda: Recognizing the crucial role of Urban Authorities in daily life, the European Urban Agenda has een created, to promote cooperation between Member States, cities, the European Commission, and other stakeholders. The agenda aims to offer a cooperation platform to enhance sustainable and efficient urban mobility, focusing, among others, on public transport, soft mobility, and accessibility for disabled and elderly individuals. Transport-Specific Policies Apart from those very general and multi-field covering regulations and recommendations, there are also two important regulations that investigate specifically transport at an international level. - White Paper on Transport: In 2011, the European Commission adopted a roadmap of 40 initiatives for the next decade to build a competitive transport system that increases mobility, removes major barriers, and fuels growth and employment European Commission and Directorate-General for Mobility and Transport [\(2011\)](#page-111-14). The objectives include removing conventionally-fueled cars in cities, promoting low-carbon fuels, and reducing greenhouse emissions in the transport sector. These goals were later refined by the Green Deal and "Fit for 55" initiatives. - European Mobility Strategy: Launched in 2021, the European Commission [\(2021c\)](#page-111-5) targets mobility with several implications for transport poverty. Firstly, it identifies the significant challenge of reducing emissions and becoming sustainable. Then, the strategy emphasizes that "fostering cohesion, reducing regional disparities as well as improving connectivity and access to the internal market for all regions, remains of strategic importance for the EU". (para.5). It also stresses, in paragraph 29, that the sustainable transition should leave no one behind. Additionally, the strategy highlights that people are willing to switch to more sustainable alternatives, with cost, availability, and speed being the main motivators. Finally, the document exposes that mobility patterns are changing due to teleworking and other factors, so this should be taken into account. It also posits that the EU is committed to promoting and incentivise the use of open data sources, indicating that it aims build a European Common Mobility Data Space. # 2.3.2.2 Spanish Policies In addition to global and high-level documents that provide international directives, each country has its own specific targets. These targets are (or should be) aligned with international directives but also tailored to the situation of each state. In particular, two elements are considered as paramount when analysing transport poverty in Spain: - LOTT and ROTT: In Spain, two main documents regulate land transport: Cortes Generales de Espa˜na [\(1987\)](#page-110-13) and Gobierno de Espa˜na [\(1990\)](#page-111-15). The first one regulates land transport in Spain, promoting free competition, safety, and service quality. The latter details specific requirements and procedures for transport operations, including responsibilities, insurance, and inspections. - Strategic Mobility Plan from the Ministry of Transportation: In 2021, the Spanish government launched the Strategic Mobility Plan from the Ministry of Transportation (Ministerio de Transportes y Movilidad Sostenible, [2021\)](#page-114-2). This strategic plan outlines the main directions the Spanish government is committed to taking. The first of the nine axes proposed by the strategic plan specifically mentions: "The 1st Axe will focus on finding solutions to make mobility accessible and affordable for all citizens, in all territories" (p.14). This demonstrates a strong commitment from the Spanish government to make mobility accessible in all regions. Additionally, the document includes the notion that: "Mobility should be considered as a right" (p.45), with all the implications this consideration entails. Adding to that, a particularly relevant sentence is: "Mobility should guarantee that citizens can cover their needs, without requiring their own vehicle" (p.62), highlighting the need for a public transport analysis rather than a private vehicle one. Apart from that, Axe 8 posits that a fair transition needs to be made in transport, helping to move towards a more egalitarian labour market. New conditions also need to be considered. For instance, in 2020, 15% of employees worked from home some days, and 10.8% did so for more than half of their week (2020, COVID-19 Pandemic year, source: INE, Encuesta de Poblaci´on Activa. Ocupados por frecuencia con la que trabajan en su domicilio particular ). Furthermore, the Spanish government aligns with the EU, emphasizing the necessity to offer competitive alternatives to fossil fuels, promote the sustainability of transport infrastructure, and encourage the use of less polluting vehicles. The Spanish government also shares the European ambitions concerning digitalization and the creation and promotion of open-source databases. • "30-Minute Country" State Action Plan: The "30-Minute Country" State Action Plan pretends to be a strategy (but is not yet in place) aimed at coordinating the actions of the General State Administration to develop a policy of territorial cohesion. Its main goal is to ensure that The plan seeks to ensure equitable access to essential services and opportunities across the territory to reduce territorial inequalities, promoting more balanced development between urban and rural areas. Based on its name, it indicates that Spain aims to become a country where all citizens can access all essential services in 30 minutes. Through collaboration between administrations, local entities, and the private sector, this model aims to ensure that all citizens, regardless of where they live, can enjoy the same rights and opportunities (para la Transici´on Ecol´ogica y el Reto Demogr´afico, [2025\)](#page-114-11). In addition to the aforementioned issues, the Spanish government is also facing a significant challenge closely related to transport: the depopulation of rural areas. To address this, the Spanish government launched the Gobierno de Espa˜na [\(2021b\)](#page-112-4) in 2021, which includes 101 measures aimed at preventing the continued decline in village populations. For instance, measure 1.18 focuses on support measures for the acquisition or replacement of vehicles to promote the adoption of plug-in hybrid and electric vehicles, thereby progressively reducing dependence on diesel vehicles in rural areas. Another example is measure 2.7, which involves designing and promoting innovative territorial connectivity proposals, such as collective transport and collaborative economy initiatives, to enhance sustainable mobility options within the framework of the "Rural Mobility Table." This measure also includes conducting a comparative European study on mobility solutions in low-density or depopulating areas. #### 2.3.2.3 Conclusion of the main directives To sum up, the considered directives provide guidance on the main direction the research should be headed towards. - Research should be focused on studying and promoting sustainable transport, prioritizing sustainable alternatives, and aiming to reduce emissions. - Research should emphasize the needs of vulnerable groups, that in the case of transport, correspond to low income classes and rural areas inhabitants. The current regulation places great importance on ensuring no one is left behind. Specific funds are allocated for this purpose, making inclusion more important than monetary issues, although efficient resource allocation should remain a priority. - People change of behaviour mainly depends on three key aspects: cost, availability and travel times. - Research should utilize and promote the creation and use of open-source data. - According to the Spanish government, public transport should ensure mobility solutions that allow citizens to meet their needs without requiring their own vehicle. - Finally, Spain is also striving towards the strategic objective of Spain becoming a "30 minute country," where all essential services are accessible to any citizen within a 30 minute time frame. # 2.4 Concluding remarks Considering this literature review, it can be stated that the definition of transport poverty is relatively well established, but it still evolving and has some important areas (specially concerning the way to measure it) where consensus has not yet been obtained. Generally this concept is addressed through four dimensions, that cover and organize a wide range of transport issues. The four dimensions are: availability, adequacy, affordability and accessibility. Currently all dimensions are considered at the same time in literature, suggesting it might be useful to organize, and structure the concepts, for a better understanding of transport poverty. Advances in this direction are presented in Section [3.1.](#page-44-1) Finally, the notion of poverty should also be present throughout the whole analysis of transport poverty. Having established these definitions and focusing on accessibility (the purpose of this work), this Chapter examines several publications in relation to the topic. These publications add different points of view concerning accessibility, and provide a good overview of different manners it has been studied so far. One of the key insights that can be taken from it is that data availability can be an issue in this kind of analysis, and numerous publications have to limit the amount of facilities considered in their studies due to lack of data. In addition, another insight relates to the complexity of the model. While complex models will output more precise results, they also require bigger amounts of data, making them less replicable and computationally intensive. Conversely, more generic indicators, are easier to compute, but may lack nuance, offering too generic results that are not actionable. Another key insight arises when analysing accessibility in the context of transport poverty versus when just looking at general accessibility analysis. In the first, it is required to define a list of essential services that every citizen should be able to access. Without such a list, it is possible to measure general accessibility and determine which areas are better or worse served, but it becomes morre difficult to identify when a situation of transport poverty exists. Looking at the three groups of articles in which the Section [2.2.1](#page-27-2) divided publications, the third group (Accessibility based on proximity to essential facilities) appears as the more suited to the purposes of this thesis. This is because the first group analyses general accessibility, but does not pay attention to the list of destinations analysed, not guaranteeing that they all are essential. As for the second type, using a mix of indicators provides indirect results, that are not that straightforward and intuitive, making them less actionable and hence suitable for decision-makers. Among all the studied works, one stands out: Radics et al., [2024.](#page-114-0) This work conceptualizes accessibility as a process that involves determining whether the travel times to essential destinations are reasonable. This perspective closely aligns with the working definition of accessibility proposed in this Chapter. Additionally, it provides a comprehensive list of essential destinations, categorized in different groups, based on an extensive literature review. Therefore, this work will serve as a starting point for this thesis, and the methodology outlined will build upon and refine the exposed methodology there, to make it appropriate for this study. Finally, the regulation review indicates that efforts should be put in creating an indicator that aligns with the objectives of the EU and Spain in the context of transport poverty. This implies to support sustainability goals, focus in the vulnerable groups and promote the use of open-data. When developing this indicator, it should offer valuable insights that are directly targeted by the law and easily linkable to it. # Chapter 3 # Methodology Based on the previous literature insights, this Chapter first tries to link the different dimensions of transport poverty, and to assemble them in a unique framework. This framework is exposed in Section [3.1.](#page-44-1) Based on it, accessibility comes as the last dimension to be analysed, and should only be measured for those travel modes alternatives classified as available, adequate and affordable, and only considering those trips that are necessary to actively participate in society, this is essential trips to essential destinations. After that, the proposed indicator for measuring accessibility is exposed, with a detailed explanation on what data is required to compute it, and which calculations are involved in the process. The methodology builds upon the work of Radics et al. [\(2024\)](#page-114-0), positing that accessibility in this context should primarily focus in measuring travel times to essential destinations. Further aligning with the European and Spanish interests, the proposed calculation method focuses on public transport, making use of open data sources. Considerations on how the parameters established can influence its adequacy to analyse rural areas are also provided. As this measurement addresses a very sensible topic, a section is also included explaining the motivations behind this indicator. Additionally, since the proposed methodology builds upon the work of Radics et al. [\(2024\)](#page-114-0), Section [3.2.5](#page-59-0) is dedicated to compare the current methodology with the one exposed in this paper. # 3.1 Link between accessibility and transport poverty Having examined the concepts of transport poverty and accessibility, it is now relevant to establish the relationship between them. Transport poverty is generally defined by four key dimensions: availability, adequacy, affordability, and accessibility. While these dimensions are inherently interconnected, structuring their order of consideration can provide a clearer framework for analysis. To this end, a method based on an individual's decision-making process is proposed, and illustrated in Figure [3.1.](#page-45-0) The workflow shown suggests a way to systematically analyse the transport network of any individual, in order to determine if the individual suffers from transport poverty. In the first place, and before starting detailing the workflow, it is reminded that this accessibility assessment is done within the framework of transport poverty. This, mainly entangles that not all trips should be considered, but only the essential trips, providing access to essential services (see Section [2.1.1\)](#page-22-2). Therefore, the list of essential services is an aspect of paramount importance in this work and is particularly relevant for the accessibility analysis. Consequently, Section [3.2.2.1](#page-52-1) dedicates significant effort to refining and determining this list. However, in this section, the main objective is to understand the link and relationship among the different dimensions, so this discussion is left for following sections, and here the list of essential destinations is assumed to be known. ![](_page_45_Figure_2.jpeg) Figure 3.1: Flowchart to evaluate if an individual is suffering from transport poverty, a proposal. Looking again at the workflow in Figure [3.1,](#page-45-0) the proposed analysis starts with an exhaustive list of every conceivable alternative of transport, which can consist of a unique mean of transport or a combination of several of them (see N in the figure). Then, the different alternatives are filtered, attending to different criteria related to three dimensions of transport poverty: availability, adequacy and affordability. Examples of possible criteria to be used for the different dimensions are offered at the right side of each decision case in the Figure [3.1.](#page-45-0) It is important to note these criteria is intended to provide some guidance and to facilitate the understanding and following of the scheme, and should not be interpreted as prescriptive standards. If the alternatives are not considered to fulfil sufficiently the requirements in any of the dimensions, the alternative is dropped for the following steps. When moving forward in the workflow, it is possible that at some point, the list becomes empty, if all alternatives have proven to fail in any of the dimensions. In this case the individual is qualified as suffering from transport poverty. Finally, the last element of the flowchart relates to accessibility. With the remaining list of alternatives (which are considered available, adequate and affordable), it is needed to analyse if all the previously established essential destinations can be reached in a way that complies with the accessibility criteria. In this case, for illustrative purposes a maximum travel times is proposed, but other requirements, for instance, requiring to arrive before a given hour could also be applied do determine the accessibility compliance parameter. If all destinations can be reached in an accessible way, using any of the remaining alternatives, it is determined that the individual does not suffer from transport poverty. Conversely, even if only one of the essential destinations does not comply with the accessibility criteria, then the individual would be characterize as suffering from transport poverty. This relation only holds true because only essential destinations are being contemplated. If non-essential facilities were contemplated, the methodology would rather be analysing general accessibility than accessibility in the context of transport poverty. Finally, it is remarked that, even if in the case an individual is classified as poor, there can also be different severities. For example, a person that can not access any essential destination may be suffering a more severe situation of transport poverty than a person that can not access one of the essential services. An example on how to theoretically apply this methodology is presented for a female individual, wishing to travel to three destinations that have been considered as essential: work, a medical centre and a supermarket. For greater clarity, Figure [3.2](#page-46-0) is presented. ![](_page_46_Figure_3.jpeg) Figure 3.2: Illustrative application of the flowchart. At the beginning, a list with all the conceivable alternatives of transport for the individual is prepared, and the analysis starts. The first factor the individual would assess is availability, determining which transport options are at her reach. She would thus leverage whether she owns (or has access to) a private vehicle or whether public transport options are sufficiently close to be considered viable. In the illustrative example, tram is discarded because the tramway stop is considered to be too far away, and also car, as the individual does not have a driving license. All other means are considered available. Once the available transport options have been identified, the individual would then evaluate their adequacy. This involves discarding options that may be perceived as unsafe, physically inaccessible due to mobility constraints, or too complex to understand sufficiently. In this case, bike is discarded because of not being considered safe. Note that the adequacy requirements are highly dependent on individuals, so this step is particularly difficult to generalize, and may explain why no indicators are proposed in Cludius et al. [\(2024\)](#page-109-0). For those transport options deemed adequate, the next consideration would be affordability. The individual would assess whether the associated costs are within her financial means. For instance, while a taxi might be an available and adequate option, in the example it would be discarded because the fare is prohibitively expensive. Only after meeting these three conditions—availability, adequacy, and affordability—would the individual consider the time required to reach each of the destinations, which directly relates to accessibility. In the example of Figure [3.2,](#page-46-0) the medical centre and the supermarket can be accessed complying with accessibility criteria, but this is not the case for work. This could be, for instance because the individuals' work location is poorly connected by public transport or is too far away. Therefore, because the individual can not access all the essential destination, it will not be able to fully participate in society, and thus can be considered to be experiencing transport poverty. This conceptualization is considered valuable as it allows for the sequencing and disaggregation of the different components contributing to transport poverty. For example, if a bus stop is located too far from a household, this would undoubtedly impact accessibility by increasing travel times. However, under this framework, it would be classified primarily as an availability issue rather than an accessibility concern. This would translate in considering that this person can not access to the network, rather than signalling that the network is too slow (too long travel times). Moreover, this structured approach aids in prioritizing interventions. For instance, it would be ineffective to focus on improving accessibility in a rural area (e.g., by reducing travel times to essential destinations) without first ensuring that the proposed transport options are available, adequate or affordable. If a transport mode is too far away, too complex, unsafe, not adapted, or too expensive, individuals will not utilize it regardless of its accessibility. Finally, this methodology also allows accessibility to be assessed for both; individuals but also target groups. Indeed, this analysis can be performed at an individual level, but also at higher levels of aggregation. By using representative profiles for each group instead of individual data, the analysis can help measure transport poverty at a group level. For example, a scenario could be created for an elderly person living in a rural village, which would act as a representative persona of this group, making it possible to infer conclusions for this whole population segment. By analysing transport poverty in each village, the data can be combined to calculate broader statistics, such as the percentage of elderly individuals facing transport poverty across the entire area. This could be done for all the target groups of interest, and would enable to differentiate who and where people need help. Thus, by sequentially evaluating transport poverty through these four dimensions, a more targeted and effective approach to policy and planning can be achieved. On the other hand, a key limitation of this approach is that it disregards personal preferences and attitudes toward transportation. In the context of transport poverty, this is considered as acceptable, since the objective is not to accurately predict actual choices people make, but to assess if there is a situation of transport poverty. However, this perspective introduces a significant weakness: the framework is somewhat impersonal and rigid. By relying on binary classification criteria, it fails to consider subjective factors that influence whether a transport option is truly a suitable option to individuals. For example one such factor is perception. Political views, cultural attitudes, or personal preconceptions about certain modes of transport can make them unacceptable to some users, even if they technically meet the framework's criteria. This subject is studied in numerous publications that investigate attitudes towards public transport (such as Beir˜ao and Sarsfield Cabral [\(2007\)](#page-109-9), Urbanek, [2021,](#page-115-7) Mann and Abraham, [2006\)](#page-113-9), and that highlight that these factors are indeed important and should be accounted for. However, because these factors are not considered in the proposed methodology, individuals in such situations would not be classified as experiencing transport poverty—even if, in practice, they are unable to travel. This reflects a fundamental assumption of the framework: transportation is a necessity and consequently people will always choose an available mode of transport, regardless of personal beliefs. Finally, it is noted that this model is only a proposal, and needs further validation and discussion, to assess whether it is convenient for all kind of scenarios, or further refinements is required. # 3.2 Proposal of an accessibility indicator Accessibility is defined (see Section [2.1.2\)](#page-25-0) as the possibility to access essential services/destinations in a reasonable amount of time. According to Figure [3.1,](#page-45-0) this analysis only takes place once the means of transport considered have proven to be available, adequate and affordable (this means there is at least one mean of transport that fulfils all the requirements). At this point, it is required to develop a methodology to compute whether this means of transport are accessible, and therefore whether or not a person suffers from transport poverty. In this context, the objective of the indicator is not to see which areas are better or worse served, but rather if there are areas that are not well served enough, implying that their citizens suffer from transport poverty. This categorization does not undermine the idea that different levels of poverty can exist, having areas more or less severally affected by transport poverty. Hence, it is first required to decide which will be the areas of interest, this is, the level of aggregation (households, districts, municipalities, provinces, countries, etc.) of interest. This choice will mainly be driven by the available data, and the level of detail that wants to be pursued by the analysts or the decision makers. Considering the literature presented in Section [2.2.1,](#page-27-2) it is evident that accessibility is not a new concept and has been explored in previous research. However, a particular challenge arises when analysing accessibility within the context of transport poverty: it is necessary to define a list of essential services, provided in essential facilities that every citizen should be able to access. Without such a list, it is possible to measure general accessibility and determine which areas are better or worse served, but it becomes impossible to identify when a situation of transport poverty exists (see Section [2.4\)](#page-42-1). Therefore, the methodology proposed bears the list of essential services as a core element, and is based in the publications of Radics et al. [\(2024\)](#page-114-0), that is taken as a foundational reference. The other publications reviewed have also been taken into account, aiming to deliver an indicator as comprehensive as possible. Finally, the proposed methodology is aimed to be insightful, concrete and relevant for decision makers. ### 3.2.1 Exposition of the suggested indicator Considering its definition (Accessibility is the possibility to access essential services in a reasonable amount of time, Section [2.1.2\)](#page-25-0), it seems reasonable to assess accessibility by first identifying which are the essential services required, and which are the facilities that provide access to them. For instance, the service of healthcare is provided in hospital facilities. Since people usually do not live at the same place where the services are provided, they will need to travel and therefore perform trips to access those facilities. These trips allowing access to essential facilities can be referred to as essential trips. After defining those trips, it is required to assign a reasonable amount of time, (this is a maximum time) in which citizens should be capable of performing these essential trips. Finally, the accessibility analysis would consist of analysing how many of these essential trips can be done within the reasonable amount of time threshold. This align very closely with the work presented in Radics et al. [\(2024\)](#page-114-0), that is used as a starting point of the proposed methodology. Section [3.2.5](#page-59-0) exposes the main similarities and differences of both methods. Yet, it seems likely that the list of essential services may not be the same for all citizens, and may be related to social aspects, requiring to differentiate the specific needs for each target group. For instance, the essential destinations for kids, may be different to those for working adults or retired citizens. Therefore, a correct indicator should be able to capture this nuances for each group. Adding to that, while it is important to be able to perform all essential trips, it is also intuitive that not all of them are done the same number of times. In that sense, it is generally observed that the most frequent trips tend to be the most important ones for individuals, as people will suffer this transport poverty (in case it exists) more often. For example, both, commuting to work and going to a hospital are essential trips each citizen should be able to perform in a reasonable amount of time. However, not reaching work in a reasonable time seems more of a problem than not reaching a hospital (in a non-emergency situation), because the individual goes to work 5 days a week, but only visits the hospital a few times a year. Consequently, the frequency of each trip can serve as a weighting factor, which assigns relative importance to each essential trip, ultimately contributing to an overall accessibility indicator. Aligning with this view, an indicator to compute the accessibility levels of an area of interest is proposed. To compute the indicator several steps are required: - 1. First, identify the K target groups (i), the list of L essential facilities (j) and the S areas of study to be analysed, for instance municipalities. Then, assign for each group i, a reasonable amount of time to reach each essential facility j (Tij ) and determine the frequency of visits per month of each group i to each essential facility j (Fij ). These values will remain the same for all S areas of study. It is noted that while the list of essential facilities is the same for all groups, the reasonable amount of time associated to them and the frequency of visit may vary across groups and across essential facilities. In addition, if the frequency of visit of a group i to an essential facility j is 0 (Fij = 0), this means that the essential facility j is not visited by group i. Finally, each group from each area of study should be awarded a start location/ domicile, that could be understood as the prototypical location in where a person from this group and this area of study would live. This technique is commonly used in traffic simulation programs (such as AIMSUN, SUMO, VISSIM, etc.), where the population of each area of study is usually assumed to arrive or depart from centroids. The ultimate goal of determining this location is to establish which means of transport can be considered as available. Up to this point all inputs depend on the analyst, the decision makers and the assumptions or facts from literature found, but no calculations are involved. - 2. Second, perform travel time calculations, and see if they comply with the required time thresholds. To that end, it is first necessary to compute, for each area of study z, the time required to reach each essential destination j (ex, hospitals, parks, schools, etc.) for each group i. This parameter is called tijz. Then, to determine if the obtained value is below the reasonable time previously established to reach facility j by citizens of group i (Tij ), this is if tijz <= Tij . If this is the case, the binary compliance indicator, Cijz, is set to 1 (Cijz = 1) indicating that it is possible for citizens of group i living in the area of study z to reach the essential facility j in a reasonable amount of time. Else Cijz = 0. It is noted that the reasonable amount of time is the same for all areas of study, but the time to reach each essential facility is different for each study area. To compute this time, accessibility techniques from literature could be used (such as all exposed in Section [2.2.1\)](#page-27-2), and additional explanations concerning these calculations are provided in Section [3.2.3.](#page-55-0) At this stage, data availability will importantly condition these measurements, and approximations or simplifications may have to be used. 3. Third, the accessibility indicator for each group i in each area of study z (Aiz) can be computed as the percentage of essential trips that can be performed in a reasonable amount of time by each group. The formula to be used is shown in Equation [3.1,](#page-50-0) which takes into account the frequency of each essential trip for each group i to each essential facility j (Fij ) and the compliance indicator described above (Cijz) to compute the total number of essential trips that can be done in a reasonable amount of time. $$A_{iz} = \frac{\sum_{j=1}^{L} F_{ij} C_{ijz}}{\sum_{j=1}^{L} F_{ij}}$$ (3.1) The result of this equation will yield a value ranging from 0 to 1 (equivalent to 0% to 100%), where 0% indicates that individuals from that group and area are unable to reach any essential facility within a reasonable amount of time, and 100% signifies that all essential trips can be completed within a reasonable time frame. It is important to note that the only acceptable threshold of accessibility should be 100%, as this is the rationale behind their designation as essential. However, in the context of transport poverty, there can also be different severities. For instance, a region that can achieve 90% of their essential trips will suffer from transport poverty, but will be less severely affected by it than another region where only 40% of the trips can be performed. Because public transport configuration is the same for all citizens, all citizens living in the same area and belonging to the same group can be assumed to have the same accessibility. Therefore, the number of residents of each group in each area of study is also considered (Piz). All these three points can be visualized in a more comprehensible manner in a table. To that end, the resulting formulation for a municipality m (considered as the area of study) is presented in Table [3.1.](#page-51-0) There, it can be seen how all the previously defined parameters and indicators are organized in columns, to provide a clearer overview of the calculation process. 4. Finally, compute the overall accessibility of each area of study (Az). This is the result of computing the weighted average of accessibility for each group i in each area of study z (Aiz), using the number of citizens of each group (Piz) in that area of study as weight. To do this, the following Equation [3.2](#page-50-1) is presented, where Az is the total accessibility for the area of study z , Aiz is the total accessibility of residents in that area of study z of group i, and Piz is the number of residents in that area of study z of group i. The interpretation of the results is identical to the one of Aiz. $$A_z = \frac{\sum_{i=1}^K A_{iz} P_{iz}}{\sum_{i=1}^K P_{iz}}$$ (3.2) Following the example in Table [3.1,](#page-51-0) if municipalities are taken as the different areas of study, the total accessibility for a municipality m (z = m) would be computed as shown in Table 3.1: Proposed indicator to measure accessibility in a municipality m for K groups and L essential facilities. | Group
(Gi) | Essential
facilities
(EFj) | Reasonable
time (Tij) | Number
of trips
per
month
(Fij
) | Does it
comply*
(Cijz=m) | Accessibility
(Aiz=m) | Number
of
residents
(Piz=m) | |---------------|----------------------------------|--------------------------|-------------------------------------------------|--------------------------------|--------------------------|--------------------------------------| | | EF1 | T11 | F11 | C11m | | | | Group
1 | EF2 | T12 | F12 | C12m | A1m | P1m | | | | | | | | | | | EFL | T1L | F1L | C1Lm | | | | | EF1 | T21 | F21 | C21m | | | | Group
2 | EF2 | T22 | F22 | C22m | A2m | P2m | | | | | | | | | | | EFL | T2L | F2L | C2Lm | | | | | | | | | | | | | EF1 | TK1 | FK1 | CK1m | | | | Group
K | EF2 | TK2 | FK2 | CK2m | AKm | PKm | | | | | | | | | | | EFL | TKL | FKL | CKLm | | | Does it comply\*: Determines whether an available, adequate, and affordable mode of transportation allows access to the essential facility within the specified reasonable time. Can be 1 or 0. Equation [3.3:](#page-51-1) $$A_m = \frac{\sum_{i=1}^{K} A_{im} P_{im}}{\sum_{i=1}^{K} P_{im}}.$$ (3.3) The suggested indicator is considered adequate, as it closely related to the definition of accessibility. Adding to that, it is also considered systematic, while generic enough, since the characteristics and differences for each group can be nuanced. Similarly, depending of the level of aggregation, the results will be more suited in urban or rural areas (a matter further discussed in Section [3.2.3.2\)](#page-57-0). Finally, in this study, accessibility is defined only as travel time to essential facilities, disregarding the time it takes to receive the service itself (as opposed to what done in Enderami et al. [\(2024\)](#page-110-14) for instance). Although the method is sought to reduce the amount of input data required, and therefore works with group, the table could also be fulfilled individually with surveys. If this was the case, it would be possible to determine precisely how many citizens from an area suffer from transport poverty and offered tailored solutions. However, the data collection process would be costly. More comments are made about the assumption and motivations in Section [3.2.4.](#page-58-0) Finally, accessibility levels will highly be influenced by the selected parameters. Therefore, decisions the adopted thresholds will have paramount influence on the results. In this regard, as signalled in Capasso Da Silva et al. [\(2019\)](#page-109-10), p.1: "accessibility is a metric, but what are acceptable parameters of what is considered accessible must be set through policy". In this study, this will not be the approach taken, and threshold will be based in literature, statistics or other sources. However the aim would be to reach a point where this matter becomes sufficiently relevant for them to be established by policy. In order to better understand this concept, a full illustrative example on how to use the scheme from Figure [3.1,](#page-45-0) and the Table [3.1](#page-51-0) is presented in the Appendix. In addition, the whole Chapter [4](#page-62-0) applies the methodology described in this Section to a concrete province in Spain, to demonstrate its applicability. ### 3.2.2 Required data for the accessibility indicator ### 3.2.2.1 List of essential facilities, frequencies and acceptable travel times As seen in Table [3.1,](#page-51-0) in order to compute the accessibility indicator proposed, it is first needed to establish the groups that want to be studied, and then for each of them which are the essential trips they need to make, the frequency of each of them and the maximum amount of time it should take to fulfil it. Guidance on how to establish these parameters is offered. Essential facilities Essential facilities are those places where essential service can be received, so it is needed to first define which is the list of essential services. However, no official list of essential services applies to the entire Spanish territory. The definition of essential services is a sensitive and often debated issue. For this reason, and although considerable effort has been made to compile a representative list of such services, it is important to acknowledge that this list is not definitive and may evolve over time. The proposed selection of essential services, and therefore facilities in this study draws upon three key sources. The first one is a 2020 presentation by Professor Carlos Moreno, available via the EIT Urban Mobility platform (Carlos and Director, [2020\)](#page-109-11), where he delves into the concept of the 15-minute city and identifies six broad categories of essential services. These categories are further refined in the work of Radics et al. [\(2024\)](#page-114-0), which includes a comprehensive literature review (see Appendix A of the publication), modifying and consolidating the list of essential services into: Living and Transport, Working, Commerce and Catering, Healthcare, Education, and Entertainment/Leisure. Due to its thorough literature review, and the similarities found between this work and the current one, it was decided to adopt this classification (see Section [2.2.1.3\)](#page-32-0). The third reference is Logan and Guikema [\(2020\)](#page-113-7). In this work, several elements that could be considered essential (and that could fall under the previously mentioned categories) are provided, but also the notion that a correct list can only be obtained through citizen participation. This work emphasizes the need for community involvement in validating and contextualizing what is considered essential. For the purposes of this study, these six categories will be used as a reference framework. However, the Transport essential service category is excluded, as accessibility to public transport stops is more appropriately analysed within the analysis of transport availability dimension, rather than under the accessibility one. Within each of the remaining five essential services, at least one essential facility shall be selected for inclusion in the accessibility indicator table. Frequencies of each trip Establishing a frequency for each trip is also a not straightforward process, however common routines can be used to obtain representative results, without requiring of numerous input data. One of the first intuition to establish this parameter would be to use general mobility surveys that contemplate the number of trips that are done by citizens in an area, for instance [EMEF survey](https://ce-sermetra.atm.cat/es/web/observatori/w/encuesta-emef) in Catalu˜na. However, this data might not be optimal for this study. This is because different types of services involve distinct usage patterns. While visits to medical consultations tend to be need-based and infrequent, only visiting them when needed (essential trips), services like supermarkets or leisure centres can be accessed more regularly than what strictly necessary. Therefore, there can be a mismatch between the number of trips that are done, and the number of trips that are essential. For instance, some people may go every day to a supermarket, but most would agree that it would be a mistake to have 7 trips a week to the supermarket labelled as essential. In that sense, general mobility surveys provide valuable behavioural insights, but may include non-essential trips to essential destinations. While no people will go to services less than what is essentially required, people may go for non-essential purposes, and in the survey data, it is not possible to differentiate them. This can distort the values needed to assess accessibility to key services. Other studies like Zhong and Bian [\(2023\)](#page-115-8) and Pappalardo and Simini [\(2018\)](#page-114-12) have also explored mobility using advanced modelling techniques, but they do not yield specific, applicable conclusions for frequency estimation within this context. A further complication lies in how people organize their trips. Travel is often multipurpose, with individuals concatenating trips into a single outing (e.g., work, sport, grocery shopping, home). Additionally, people may travel not for their own needs but as caregivers. This would be the case for instance, for parents accompanying children to school or individuals escorting relatives to appointments. These behaviours fall outside the scope of this thesis but highlight the complexity involved in defining the frequency of visiting essential destinations. In the absence of standardized datasets that clearly distinguish trip purposes, provisional values must be adopted. Therefore, the recommendation for completing this methodological section is as follows: For facilities where it is evident that individuals visit only when necessary (such as educational institutions, healthcare centres, and workplaces), utilize data from general mobility sources or relevant institutional data. For other types of facilities, which may be visited more than what is strictly essential, seek first recommended minimum access values from existing literature. In the absence of such data, a survey may be conducted within the study area to ascertain the frequency of trips per unit of time. If conducting a survey is not feasible, reasonable estimates can be inputted. Finally, note that frequency also has a major impact when differentiating groups. For instance, establishing that the frequency of visits to a facility is 0 for a group, effectively excludes considering that facility in that group analysis. Similarly, frequency is a measure that is capable of capturing different mobility patterns, allowing for differentiations within groups that may visit the same facilities, but not visit them as often. Reasonable amount of time Concerning the reasonable amount of time, this is also an unsettled matter. Several recent publications are exploring the concept of a 15-minute city (FMC), whereby residents can fulfil their daily needs and activities within 15 min of walking or cycling (Pozoukidou and Chatziyiannaki, [2021\)](#page-114-13). These daily needs are fulfilled in facilities such as "healthcare institutions, preschools and schools, shops, leisure areas, cultural and entertainment amenities, parks and natural areas" (Gil Sol´a and Vilhelmson, [2019,](#page-111-16) p.3). However, this concept only seems to be conceivable in densely populated areas, where all these facilities would encounter sufficient demand. Questions nevertheless remain on whether this 15 minute city concept could also be applicable in rural areas, and particularly in poorly populated ones. If the 15 minute concept was applied strictly in rural areas in Spain, this would result in an unreasonably large amount of schools, medical centres, etc., that would be difficult to fund and justify due to their low occupancy rates. It is the opposite phenomena which is currently observed and predicted, where schools and public institutions are expected to close due to a lack of students in rural areas (Gobierno de Espa˜na, [2021a.](#page-111-8)) Adding to that, studies in the field suggest that people living in rural areas tend to have a higher tolerance to travelling more time than those living in rural areas (McQuaid, [2009\)](#page-113-10), suggesting that higher thresholds should be considered. All these are matters that remain largely unexplored, but that are relevant for this study. Nevertheless, a robust reference has been identified to establish a threshold. Spain is striving to become a 30-minute country, as articulated by Francesc Xavier Boya, the General Secretary for the Demographic Challenge in Spain, in an interview with La Vanguardia (Bosch, [2025\)](#page-109-12). This 30 minutes threshold is also the one used by the EU Commission report (Cludius et al., [2024\)](#page-109-0) in their unique accessibility indicator. Consequently, this threshold will be applied across all categories. However, tailored thresholds could have been applied for each group, further nuancing the compliance indicator for each group. Having determined all this parameters, it would be possible to move forward in the accessibility indicator calculation. #### 3.2.2.2 Other raw data required To continue with the accessibility calculation, and in particular to fulfil the Does it comply and the Number of residents columns, the following raw data required is: - Demographic data from census tracks: This data allows to understand and bear in mind how many residents are affected by a given accessibility level. As all groups have similar characteristics and live in the same area, all of them can be assumed to have the same accessibility levels. - GIS coordinates of each area of study: Having the coordinates of the areas (polygons) under study will allow to output visual maps of the indicator, making it much easier to understand. - GTFS public transport data or network information: this data is necessary to compute travel times. - Coordinates of essential facilities: this data is required to determine which areas provide access to essential facilities. - GIS coordinates of representative locations for each group if required/available: This data is not strictly required, but can allow to apply additional filters when performing the previous availability analysis. For instance, if it is known that people of a given group concentrate around some areas (example youngsters concentrating around universities), only stops that are close-by this area could be considered. With this information, it would be possible to compute the column and finish the indicator. The computation method is explained in the following Section [3.2.3.](#page-55-0) ### 3.2.3 Calculation of the travel times and its consequences on the indicator The objective of this measurement is not to analyse actual travel choices made by individuals—an approach typically suited for traffic simulation—but rather to determine whether a destination can be reached within an acceptable time-frame. As a reminder, according to Figure [3.1,](#page-45-0) this should only be done for the alternatives of transport considered available, adequate and affordable. Therefore, a main distinction needs to be done when deciding which facilities to consider. Accessibility to facilities that centralize the provision of a service For those services that are centralized in concrete destinations (such as hospitals that centralize health services, or schools, that centralize education services), the analysis is straightforward. It consists on identifying the origin area (the area of interest) and the destination (area containing the essential service) and computing the travel using the appropriate mode of transport. Accessibility to facilities that do not concentrate the provision of a service For those services that are not directly linked to some precise location (such as job locations, that can be spread in an area), three main approaches can be taken. a) The first option for assessing this indicator would be through direct surveys, collecting data on workplace locations and commuting times. This would allow a direct comparison between actual travel times and the established threshold, but would be costly. b) The second option involves simplifying the problem by selecting a few known destinations that concentrate the service examined, using them as proxies for all the facilities where the service is provided. For instance, considering a central business district (CBD) as a proxy for all destinations that provide work, even though not all citizens work in the CBD. c) The third option, more complicated but also more precise than the previous one is to estimate a gravity model, as explained in Section [2.2.2.](#page-34-0) This involves the following steps: (1) Identifying the number of trips generated and attracted by each area under study. (2) Allocating these trips to estimate the average commuting time. (3) Computing the average travel time and comparing it to the established threshold. In addition, this thesis places special attention in public transport to compute accessibility. While car accessibility is a broadly studied matter, where important datasets such as Eurostat [\(2025\)](#page-111-0) or Instituto Geogr´afico Nacional (IGN) [\(2022\)](#page-112-8) provide extensive data on travel times by car, the same level of information is not there yet for public transport. This is why the following section investigates this questions, and proposes a method that can be computed using GTFS data, a format used by more than 100 countries and more than 10,000 agencies (GTFS.org, [n.d.\)](#page-112-9). It is noted however that the methodogy exposed above could also be computed using car information. ### 3.2.3.1 Computation method of travel times for public transport The proposed computation method is based in the use of open source GTFS data to compute travel times using public transport. This method is therefore aligned with the Spanish and European directives, as it focuses on public transport and promotes the use of open data (see Section [2.4\)](#page-42-1). General Transit Feed Specification (GTFS) is a widely used open data format that includes transport-related information. GTFS files provide public transport specifications, such as routes, schedules, means of transport (bus, train, etc.), stops locations, and agencies. Although it is an open database, and therefore ultimately depends on the information that is willingly updated by the transport operators, GTFS are growing in popularity and availability, being now utilized by over 10,000 agencies in more than 100 countries (GTFS.org, [n.d.\)](#page-112-9). Because of its availability and broad usage, GTFS data was chosen for conducting accessibility analysis. However, it is noted that this data is not always revised, so previous data treatment is crucial before starting calculations to ensure coherent results. Common errors in GTFS data can include, for example, incorrect coordinates of stops or times exceeding 24 hours. The main concern raised in this regard is that they might be incomplete, and thus not represent adequately the current situation. In that sense, it is important that administration and governments stress the importance of having this data, in order to get an image as close as possible to reality, and invest in creating an organism in charge of supervising that the provided data is correct. Using GTFS data, travel times between origin and destination stops can be computed. Below is a brief explanation of how this process is done within this project. The purpose of this section is to conceptually understand the code, but not to enter in detail in the code explanation. The full code, written in Python, is available through a [GitHub repository,](https://github.com/Kilitho/Access-request) including the required data and functions and with explanations allowing to understand what each function does. To facilitate the understanding of the computation algorithm used, Figure [3.3](#page-57-1) is presented. Note that, in order to make the whole explanation more comprehensible, the granularity of areas of study selected here were municipalities. However, every granularity level could be adopted, and municipality and areas of study can be used as synonyms in this section. Baring this in mind, to travel times between two municipalities of interest can be computed by: - 1. Establish the municipality of interest/study (origin) and find the n-closest municipality(ies) that have facilities providing the essential service of study (in the example of the figure a hospital, as healthcare service is analysed). These municipalities hosting the essential services are taken as destinations. - 2. Create two lists containing all stops in the municipalities of interest. One for the origin and one for the destination. - 3. Select one stop from each list and evaluate which routes go through them. - 4. Check if there is any route going through both stops. If it is the case, then a direct route exists. For example, in Figure [3.1](#page-45-0) route R3 connects directly stops A2 and C1. - (a) If a direct route exists, compute the travel time between the stops using GTFS schedules (arrival time to end municipality minus departure time from origin municipality), ensuring the arrival time is higher than departure time (this means that the line goes in the right direction). - (b) After computing this for all possible combinations of stops pairs, select the option with the minimum travel time. - 5. Calculate the travel times on indirect routes. To that end: ![](_page_57_Figure_1.jpeg) Figure 3.3: Illustration of the methodology for computing travel times. - (a) For each pair of stops (one in the origin municipality and one in the destination municipality), evaluate which routes go through each of them. Store them in two different lists. - (b) From these new lists, select a pair of routes (one going through the stop in the origin municipality and one going through the end municipality) and generate, for each, a list with all the stops the routes go through. - (c) Compare both lists to find common stops. For example, looking at Figure [3.1](#page-45-0) it is possible to see that routes 4 and 5 have a common stop: B4. - (d) In case there are common stops, check if schedules are compatible for transfers (this means the departure time from the transfer stop of the second route is greater than the arrival time to the transfer stop of first route). If so, compute the total travel time as the arrival time at the destination minus the departure time from the origin stop. The transfer time can also be constrained, to have some minimum or maximum values. - (e) Repeat for all combinations of stops and its respective routes and select the route with the minimum travel time. - 6. Select the overall minimum travel time for all combinations. #### 3.2.3.2 Implications of the computation method These methodology implies five considerations. The first and foremost implication of this computation method is that the results will only provide insights on public transport. Secondly, by following this scheme, it is only possible to compute travel times between origins and destination with direct routes or routes with one transfer. It is not possible to have more than one transfer, whcih could be a severe limitation if considering urban areas. The third consideration is that this methodology considers all stops in a municipality having an essential facility as providing access to it. Yet, this may be seen as unrealistic or not correct. However, this area of partition could be reduced and refined if the coordinates of the essential facilities are known. For instance, one could consider only the stops that are within a given distance from the essential facility. Alternatively, district partitions could be used instead of municipal partitions, if there is available GIS data. A broader discussion on this topic is offered in Section [5.2.3.](#page-94-0) A fourth important consideration is the computational intensity of the process. The runtime largely depends on several factors: the number of municipalities/areas of study included in the study, the number of destinations considered per origin, the number of stops within each municipality, and—when calculating indirect routes—the number of routes through each stop and the number of shared stops that could serve as transfer points. In this regard, calculating indirect travel times is significantly more time-consuming than computing direct travel times, and this should be taken into account during implementation. It is also worth noting that the code developed for this study has been optimized only within the scope and constraints of this work; however, there is clear potential for further refinement to improve computational efficiency. The fifth consideration is that this methodology could be used in both, rural and urban areas. However, differences in the way to proceed appear. When studying urban areas, that tend to have more population and have bigger municipalities, smaller partitions (such as districts or neighbourhoods) should be used. In addition, more data is required to properly analyse urban areas, as they cal also have urban transportation options (metro, tram or urban buses) that are usually not present in rural areas. On the other hand, if focus is paid to rural areas, higher levels of aggregations can be used, and less GTFS data is required, decreasing significantly the computational time. However, less precise results will be obtained as the movements within the urban areas where services facilities are usually concentrated will not be taken into account. The code is available through a [GitHub repository](https://github.com/Kilitho/Access-request) and in the submission in Canvas. ## 3.2.4 Assumptions and motivations of the indicator Looking at the previous table, the proposed indicator may appear overly rigid and lacking nuance. However, this topic is highly sensitive, as it directly impacts social inclusion and significantly affects people's lives. Therefore, it is relevant to clarify the rationale behind these choices. • The aim of this methodology is to provide an operative framework and to output numerical results: Because this study aims to provide quantitative results, simplifications had to be made. Therefore, this methodology might not be reflecting correctly some situations, in which some users may be suffering from transport poverty, and that could not be identified in this methodology. - The indicator is dichotomous but not the thresholds: while it is true that this methodology classifies individual as suffering or not from transport poverty, it is considered generic enough to accommodate to various situations. For instance, instead of assuming a universal threshold of 30 minutes, these thresholds could be refined and adjusted for each group, better reflecting their travel times tolerance. In addition, while transport poverty is considered a yes or no question, it is recognized that there can be different severities, that will be reflected in the accessibility measurements. This choice of addressing transport poverty as a black-or-white issue, lies in the will to make this an indicator easy to use and understand by decision makers. In addition, the checklist approach adopted aligns with commonly used procedures to determine eligibility for financial aids. Such processes typically involve verifying a set of predefined criteria, and only if all conditions are met does the applicant qualify for support. This indicator follows a similar logical framework. - The indicator tries to be comprehensive and incorporate key lessons from previous publications: on the one hand, the proposed indicator is believed to be comprehensive as it is only computed for the means considered available, adequate and affordable, thus ensuing that all other dimensions of transport poverty are considered. On the other hand, as expressed in the previous Section [2.2.1,](#page-27-2) this indicator builds upon and incorporates lessons learned with previously existing literature. - The indicator suggests the participation of the public: As signalled in Logan and Guikema [\(2020\)](#page-113-7), it is paramount to consult the local community in order to detail the list of essential services. Similarly, public populations should also be integrated when defining the frequency of visit of each service, and to determine the compliance criteria for all the dimensions, particularly in relation to adequacy - Transport is inherently political. Although this indicator aims to be objective and aligned with accessibility principles, transport remains a policy-driven issue. Decisions regarding population segmentation, essential service and facilities definitions, and reasonable travel times will always involve a degree of subjectivity. It is ultimately the role of decision-makers to determine how strict these requirements should be. # 3.2.5 Comparison of the proposed methodology with Radics et al. [\(2024\)](#page-114-0). As previously stated, this methodology builds upon the methodology exposed in Radics et al. [\(2024\)](#page-114-0), so it is relevant to clarify what are the similarities and differences with this work. In the first place, this method adopts the list of essential services (called "Destinations categories" in the publication), and the idea of establishing a reasonable amount of time to reach each essential facility. However, several modifications are also included. Namely: - The main modification of the methodology consists in adding a frequency parameter, which acts as proxy of the relative importance of reaching each facility. Therefore, those facilities more frequently visited hold a greater weight when measuring total accessibility than those that have a lower frequency of visit. - The second major change consists of adding total population of each area of study next to the accessibility measurement, not to loose sight of how many citizens are affected by a potentially low accessibility. - Third, the calculation of travel times is different, as it considers public transport travel times, and not active mobility measurements. In future steps, both methodologies could be assembled to provide more accurate and comprehensive results. - Fourth, the proposed methodology also accounts for the variability of essential services across different groups present in society, such as elderly, youngsters, etc. This allows for a more flexible method, that offers lower levels of granularity. - Finally, this work presents accessibility as the percentage of trips that can be done in a reasonable amount of time, rather that as a summation of all the facilities that can be reached in a given time. This provides, to the authors view, a better perspective to assess transport poverty, while the second indicator would be more suited when talking about resilient cities (as done in Logan and Guikema [\(2020\)](#page-113-7)). To sum up, the current indicator builds upon the existing methodology, but adds a few variation making its results more actionable, easier to interpret and corresponding precisely to the transport poverty definition. This is possible because of the addition of a frequency parameter, that allows to provide different importance to the different essential facilities, and to differentiate among the needs of different groups. The computation method is also adapted to public transport. # 3.3 Conclusions In this section, significant effort has been put aiming to disentangle and compartmentalize the various dimensions of transport poverty, as illustrated in Figure [2.1.](#page-25-1) According to this framework, the concept of accessibility becomes relevant only when the other three transport poverty requirements are already met: meaning there is availability, adequacy, and affordability of transport. In other words, accessibility should be assessed only if a available, adequate and affordable alternatives of transport exists for the population in question. Building on this understanding, a methodology is proposed to evaluate accessibility, which involves a series of decisions and parameter settings. First, the areas of study (level of aggregation) must be decided, and population must be divided into relevant social or demographic groups, depending in part on the granularity of the available census data, and in part of the decision makers' interests. Next, a list of essential facilities (and therefore of essential services) to be included in the analysis is established. For each service, two parameters are defined: the frequency with which the service is typically accessed (e.g., number of visits per month), and the maximum reasonable travel time to reach it. Once these parameters are defined, travel times can be computed using the approach described in Section [3.2.3.](#page-55-0) While the proposed method is thorough, simplifications can be applied when necessary to reduce computational demand. Similarly, the model can also be refined to include more detailed analysis. The final step involves comparing calculated travel times to the predefined thresholds to determine which services are accessible within reasonable time limits, thereby deriving the total accessibility score. Although this methodology provides a structured approach to measuring accessibility within the broader framework of transport poverty, it involves multiple subjective decisions that can be debatable. These include choices about area level of granularity, population groupings, service selection, and travel time thresholds. It is expected that a certain consensus could be achieved in the near future, but to the best of the authors' knowledge, is not there yet. Therefore, efforts have been made to justify and document the rationale behind each choice. Still, the aim of this work is not to establish a definitive model, but rather to demonstrate how such an indicator could be applied and to offer a new lens through which to examine transport poverty. Thus, the decisions presented should be seen as reasoned examples —intended to inspire further refinement and adaptation— rather than prescriptive standards. Finally, this methodology refines the one proposed in Radics et al. [\(2024\)](#page-114-0), adapting it to public transport and to transport poverty. # Chapter 4 # Results. A case study in Tarragona Following the exposition of the suggested indicator in the previous Chapter, a case study is conducted in a specific region of Spain to prove that the proposed methodology can be operationalized, and evaluate the results it could output. To that end, this Chapter first exposes all the decisions that need to be made before computing the indicator. Then, the results obtained are discussed. Even if not fully realistic due to some simplifications made, the results allow to see which kind of results and the level of detail that could be obtained if applying the methodology. In this section, several choices have been made to ensure that the results of this study are aligned with the directives of the European Union and the aspirations of the Spanish government. These choices have led to results that are suited for rural (rather than urban) areas, and the use of public transport, not private vehicles. The focus on rural areas stems from the recognition that their citizens are considered a vulnerable group in the context of transport poverty (Cludius et al., [2024\)](#page-109-0), and that "no one should be left behind" (European Commission, [2021c\)](#page-111-5). The emphasis on analysing public transport is primarily driven to align with the Spanish directives, stating that every citizen should be able to reach services without having to make use of a private vehicle (Ministerio de Transportes y Movilidad Sostenible, [2021\)](#page-114-2), and by the ambition to promote this more sustainable mode of transportation compared to private cars (Ritchie, [2023\)](#page-114-3). # 4.1 Tarragona, a region in Catalu˜na, Spain The objective of this section is to operationalize the methodology exposed above, and prove that it is applicable based in real data. The method is general, and can be applied in a broad scope. However, this thesis emphasizes that the results of it should be actionable and of direct use for decision makers, and therefore the selection of the study case was done in this direction. It was decided to focus in rural areas, perceived as a vulnerable group when talking about transport poverty (Cludius et al., [2024\)](#page-109-0) and to focus in public transport, aligning with the Spanish mobility strategy stating "Mobility should guarantee that citizens can cover their needs, without requiring their own vehicle" (Ministerio de Transportes y Movilidad Sostenible, [2021,](#page-114-2) p.62). The focus in rural areas has consequences for the calculation of travel times, that are mainly a consequence of the different network transportations present in urban and rural areas. While urban areas inhabitants have to move within urban areas to access essential services, making use of transport network such as tram, metro, active mobility or urban buses, rural areas residents need to move to other areas to access those services. Therefore, the networks to include are significantly different. A broader discussion onthis topic is provided in Section [4.2.5.](#page-72-0) Given the significance of Spanish rural areas, which represent more than 80% of the territory and are home to over 7.5 million people, it seems that they could present an interesting case study. Adding this to the fact that the study is being conducted within a company based in Spain, it was deemed adequate to perform a case study in the Spanish territory. Spain is divided into autonomous communities, each with numerous competences, allowing them to act as independent yet interconnected entities. Each of these communities are again subdivided in one or more administrative regions called provinces, that have inside each of them municipalities, the lower level of entities that is common to the whole territory. In addition, in Catalu˜na there is an intermediate level, Comarcas, which group several municipalities, but is not as big as a province. Each Comarca has a capital (the regional head) that acts as an administrative centre. Some municipalities also have lower levels of disaggregation, such as districts or neighbourhoods. For instance Madrid, the capital of Spain, has 21 districts, that are subdivided into 131 neighbourhoods [\(Subdivisions of Madrid\)](https://www.madrid.es/portales/munimadrid/es/Inicio/El-Ayuntamiento/Estadistica/Areas-de-informacion-estadistica/Territorio-y-medio-ambiente/Territorio/Mapas-de-Distritos-y-Barrios/?vgnextoid=240d64c49579f410VgnVCM1000000b205a0aRCRD&vgnextchannel=e59b40ebd232a210VgnVCM1000000b205a0aRCRD). Initially, the intention was to conduct the study in one of the autonomous community of Spain. This would be the optimal scope, as most of the transport competences are delegated to this entity level, and therefore most transport related decisions are taken by the autonomous community representative members. For instance, it is usually the autonomous community who is in charge of planning the buses that connect the different municipalities (interurban buses) in their territory (even if these competences can also be assumed by other entities such as consortiums). Therefore, Catalu˜na was selected due to its high population (7.5 million citizens), its economic relevance (19 % of the Spanish GDP in 2023, according to Instituto Nacional de Estad´ıstica [\(2025\)](#page-112-10)) and to its abundant publicly available data. Indeed, during recent years, Catalu˜na has made great efforts to digitalize its data, and now has a wide range of publicly accessible datasets, including thorough GTFS transport data, among others. However, due to the existing limitations- mainly related to the computational power available the scope was adjusted to focus solely on the province of Tarragona, one of the four provinces of Catalu˜na. This province has a population of approximately 850,000 inhabitants, representing 1.73% of the Spanish population, and comprises 184 municipalities. Based on the Spanish classification (Cortes Generales de Espa˜na, [2007\)](#page-110-6) a municipality is considered rural if it has fewer than 30,000 inhabitants and a population density below 100 inhabitants per square kilometer). Considering this, 125 of the 184 municipalities in Tarragona are considered rural areas (68%). Given that this study is conducted in Spain and targets decision makers, it is considered that the most appropriate way to divide the territory is the one established by the Spanish government, based on municipalities, as opposed to the rural/urban classification of the EU, that is based in squared grids ,that are then translated to LAUs (Local Administrative Units)(Eurostat, [2021\)](#page-111-12). This choice was made to facilitate the classification into rural and urban areas, to make the indicator easier to understand and to award results that are more tangible and relatable to general public. By taking this scope, it should be easy for decision makers to locate the problem, to conceptualize it, and to determine which is the more suited entity to solve it. For instance, given a municipality, it is possible to directly know who is in charge of the inter-municipal bus connexions that connects it (the autonomous community government, the consortium, etc.) offering a clear target for the information. However, this level of aggregation, also comes with some implications in the results, that are discussed in the following Section [4.2.5,](#page-72-0) and that mainly entangle that results will be more suited for rural than for urban areas. To facilitate the discussion of the results, Figure [4.1](#page-64-0) includes the names and locations of the most populated municipalities in the region, the regional heads, and some additional municipalities that will facilitate the following discussion. ![](_page_64_Figure_3.jpeg) Figure 4.1: Names of several municipalities of Tarragona Province. Regarding the population, Figure [4.2](#page-65-2) shows that the population of Tarragona is far from being evenly distributed. Almost all the population is concentrated in municipalities that are close to the sea (south-east). In contrast, there are numerous municipalities, particularly in the northern area, that are almost empty. ![](_page_65_Figure_1.jpeg) Figure 4.2: Population per municipality in Tarragona according to [INE Census 2024](https://www.ine.es/jaxiT3/Tabla.htm?t=68065&L=0) data. # 4.2 Establishment of the different parameters required for the measurement As outlined in the methodology description (see Section [3.2.2\)](#page-52-0), the following parameters must be established to compute the accessibility indicator: target groups, essential facilities (providing access to essential services), frequencies of essential trips, maximum reasonable travel times, census data, as well as the level of aggregation and the transportation network to be included, which will directly relate to the computation of the travel times. The decisions taken for this case study are exposed below. ### 4.2.1 Definition of the target groups Three main target groups have been identified and will be used to assess accessibility using the proposed indicator. Since the indicator requires knowing how many individuals belong to each group (to complete the column Number of residents, see Table [3.1\)](#page-51-0), the classification is constrained by the available census data, which provides age-disaggregated information in five-year intervals at the municipal level for 2024 (see [INE Census 2024\)](https://ine.es/jaxiT3/Tabla.htm?t=68535). The groups have been classified consequently. Hence, the three target groups are: • Students (ages 10–19): This group captures adolescents. For illustrative purposes, it is represented by a 15-year-old girl, which would be therefore attending compulsory secondary education (Educaci´on Secundaria Obligatoria) in Spain, which covers ages 13 to 16. In Spain, most students continue their studies until the age of 18, often in the same establishment, therefore this choice should be representative for most of the citizens in this group. Children under 10 are excluded, as they are less likely to travel independently or use public transport. Ideally, if the census data available was not in 5 by 5 years intervals, this group could have been further refined only to include the appropriate number of citizens. - Adults (ages 20–64): This group represents the working-age population and forms the core of the active labour force. - Retired (65 and older): This group includes individuals who are no longer active in the labour market. Their mobility needs may differ significantly from the other groups, and they represent every day a bigger percentage of the total population. The classification of transport needs by age group is based on the understanding that each group has distinct requirements. Students primarily travel for educational purposes, while adults' trips are mostly work-related. Elderly individuals generally have fewer essential trips as they do not work or study. This age group differentiation is evident in public transport subscriptions, where discounts are tailored to these groups, reflecting their different needs and financial means. For example[,RENFE](https://www.renfe.com/es/es/viajar/prepara-tu-viaje/descuentos) (the spanish train operator) offers specific discounts and subscriptions for youngsters and elderly, that are adapted top their different budget, but also to their different travel behaviour. This classification is also seen in other transport operators like the Madrid and Barcelona consortiums, confirming that this is a generalized and established procedure. Societal rules also support the decision of separating these groups by age. For instance, while adults can either drive or take public transport, youngsters are fully dependent on public transport, as they can not drive a car until they are 18, according to the Spanish law. Finally. this age-group classification is also endorsed by other mobility studies. For instance, a recent publication of the French mobility observatory (Union des Transports Publics et Ferroviaires (UTPF), [2024\)](#page-115-9), supports this age-based classification, proposing a Table that profiles different transport attitudes according to the generation citizens belong to, and highlighting varying attitudes and needs across generations. ### 4.2.2 Definition of essential facilities Based on the five categories of essential services to be included as exposed in Section [2.1.2,](#page-25-0) ( Working, Commerce and Catering, Healthcare, Education, and Entertainment/Leisure), the following facilities have been considered in the study, and are presented in Table [4.1.](#page-67-0) Note that for the Education category, Bachillerato1 and University1 will not be ultimately considered since the chosen representative of this group was set as a 15 years old student, attending the Secondary school. This none inclusion is reflected in the assigned frequency of 0 visits a month (see Table [4.2\)](#page-70-2) As for the supermarkets2 , filtering was applied to increase reliability and only those with a link or labelled as one of the 10 main brands of supermarkets in Spain were used. Table 4.1: Table of essential facilities and the sources of their location data for the case study. | Essential
service | Essential
facilities | Comments | Location
data source | | |----------------------|------------------------------|------------------------------------|-------------------------|--| | | | Mandatory education for | | | | | | students aged 13 to 16. In 2023, | | | | | Secondary | 81.6% of the population | GENCAT | | | Education | schools | completed this stage successfully | | | | | | (Gobierno de Espa˜na,
2024). | | | | | | Non-compulsory secondary | | | | | | education. 42.1% of the Spanish | | | | | Bachillerato | population attained a level of | GENCAT | | | | schools1 | education above the compulsory | | | | | | minimum (Instituto Nacional de | | | | | | Estad´ıstica (INE),
2024). | | | | | | Highest level of education, | | | | | Universities1 | obtained by 33.4% of the | Ministry of | | | | | population (Instituto Nacional | Universities | | | | | de Estad´ıstica (INE),
2024). | | | | Commerce | Supermarkets are included as | | OSM (filtered | | | and | Supermarkets2 | key facilities for food access and | for accuracy)∗ | | | Catering | | basic shopping needs. | | | | | Regional heads | The regional head can serve as a | | | | Working | (Cabecera | good proxy for an employment | ICGC | | | | comarcal) | hub for surrounding areas. | | | | | Primary care | CAPs provide general | | | | | centers (CAP) | practitioner services and basic | citaprevia cap | | | Healthcare | | healthcare in Spain. | | | | | | Hospitals handle emergencies, | | | | | Hospitals | complex diagnostics, and | Ministry of | | | | | specialized treatment not | Health | | | | | available at CAPs. | | | | | | Sports centres offer a wide range | Instituto | | | | Sports centres | of sport activities, essential to | Geogr´afico | | | | | maintain good health. | Nacional | | | | | Considered as representative | Instituto | | | Leisure | Shopping malls | leisure spaces, as they typically | Geogr´afico | | | | | include cinemas, restaurants, | Nacional | | | | | and caf´es. | | | As for the coordinates of the locations of these facilities, in most cases the information about the precise coordinates was available, but in some of them (namely work), only the municipality could be assigned. In any case, because the analysis is done at a municipal level, for each essential facility considered a list with the municipalities that had at least one of them is outputted. In order to enable the plotting of the results, and to assign to each point a municipality, the study utilizes a list of polygons for all municipalities in Spain, which can be obtained from [IGN.](https://centrodedescargas.cnig.es/CentroDescargas/resultados-busqueda) Some not included yet relevant destinations are: - Public transport stops: not included because they are considered being part of the availability analysis (see Figure [2.1\)](#page-25-1). - Primary schools: Because the analysis is done for public transportation (PT), it seems unlikely that young children would use it. - Police offices and stations: Although it is important to have accessible police stations, the frequency of visiting them is very low. Therefore, it was deemed unnecessary to include them in the analysis, as it would barely influence the indicator, but add a high amount of computational time. In addition, police can also go to the residence of the user if strictly necessary. - Administrative centres: in Spain, significant effort has been placed in digitalization of the administrative procedured. With the Certificado digital, an online certificate that allows to access public administration sites (similar to [DigiD\)](https://www.digid.nl/), it is possible to perform most of the online procedures remotely, making trips to the public administration less frequent. Consequently, it was thus decided to leave out of the study these places. If they were to be included, they could be placed with the work analysis, as usually the Cabeceras comarcales are the administrative centres of their regions. - Parks: In this work, the chosen level of aggregation is per municipality. Therefore, access to parks (mainly a problem in urban areas) could not be properly addressed. In addition, in rural areas access to nature is not usually an issue. - Restaurants: Although it is common for most citizens to visit restaurants with some frequency, this service has been considered to be included in the Leisure category. In addition, trips to restaurants are arguably non essential. - Places of worship: Although essential for some populations, worship centres tend to be present in most areas. In addition including them as essential trips may raise institutional concerns, as Spain officially has non state religion. - Pharmacies: This group of facilities were not included, as they are proxyed with the CAPs centers. When going to the CAP, it is usually possible to buy at a close distance the medication needed. ### 4.2.3 Definition of frequency of essential trips Determining how frequently essential facilities should be accessed, an important measure to compute the total accessibility measure, remains an open and complex issue. Yet, it is needed to consider it for the indicator, so the following monthly frequencies have been defined. The method followed is directly the one exposed in Section [3.2.2.1.](#page-52-1) This means that: for facilities where it is evident that individuals visit only when necessary (such as educational institutions, healthcare centres, and workplaces), utilize data from general mobility sources or relevant institutional data if available. For other types of facilities, seek recommended minimum access values from existing literature. In the absence of such data, a survey may be conducted within the study area to ascertain the frequency of trips per unit of time. If conducting a survey is not feasible, reasonable estimates can be inputted. For work and education, a frequency of 20 visits per month is assumed, corresponding to a five-day work or school week. The first value is a convention that is common in numerous agreements in Spain. For example, the public administration establishes it directly in Secretar´ıa de Estado de Funci´on P´ublica [\(2019\)](#page-114-14), and surveys posit that in 2022 64.3 % of the workers reported not having worked any saturday, and 78.5% reported not having worked on sunday, source: [Encuesta de poblaci´on activa.](https://ine.es/dynt3/inebase/index.htm?padre=6744&capsel=6746) This assumes that there is no home working, which could reduce the frequency of going to work. As for the education, the official [scholar calendar](https://educacio.gencat.cat/ca/arees-actuacio/centres-serveis-educatius/centres/calendari-escolar/curs-2024-2025/) establishes 5 days a week of courses during the teaching period. For sports activities, the World Health Organization recommends that adolescents engage in vigorous physical activity at least three times per week, and adults aim for 150 minutes of moderate or 75 minutes of intense exercise per week (World Health Organization, [2020\)](#page-115-10). Based on this, monthly visits to municipal sports centres are set at 12 for adolescents, 8 for adults, and 4 for elderly citizens. For this last group, the number was imputed under the assumption that some physical activity takes place outside formal facilities. For the case of healthcare, [Eurostat](https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Healthcare_activities_statistics_-_consultations&stable=0&redirect=no) reports that Spanish citizens attend doctor consultations an average of 5.6 times per year, a value corroborated by data from the [INE Health](https://www.ine.es/jaxiT3/Datos.htm?t=6369) [Dataset.](https://www.ine.es/jaxiT3/Datos.htm?t=6369) Furthermore, using [INE data on the percentage of citizens who visited healthcare](https://ine.es/jaxi/Datos.htm?path=/t00/mujeres_hombres/tablas_1/l0/&file=d05001.px) [services in the past month,](https://ine.es/jaxi/Datos.htm?path=/t00/mujeres_hombres/tablas_1/l0/&file=d05001.px) it was possible to estimate age-specific frequencies. Similarly, in Table 5.2 of Ministerio de Sanidad [\(2021\)](#page-114-15), it can be seen that in 2021, for every 1965 visit to the generic doctor, there were 484 visits to hospitals. Applying this ratio, it is possible to obtain the frequency of visiting hospitals. These frequencies were directly adopted, as it is assumed that people go to health facilities only when needed, and therefore all are essential trips. Concerning the shopping malls, no frequency information was found, so it was decided to include one trip per week (every weekend). For the elderly, the values were slightly adjusted, considering that usually this population tends to stay more at home. Hence, the values of frequency considered are presented in Table [4.2.](#page-70-2) Note1 that a frequency of 0 is assigned to Bachillerato and University, which is equivalent to not include them in the overall study. This is because of the representative persona chosen. It is again highlighted that these values are considered provisional, and if the indicator proves useful, it would be needed to refine them, using more precise and context-specific behavioural evidence. With this setting of parameters, the essential trips are distributed as shown in Figure [4.3.](#page-70-1) Youngsters and adults do a total of 36 essential trips a month, while elderly do 12. As can be seen in Figure [4.3,](#page-70-1) healthcare represents a bigger percentage of the trips for elderly than for the other categories. Work and education represent around 50% of the essential trips for adults and youngsters respectively, and sports holds significant importance for all groups. In the context of rural areas, this entangles, for instance, that proximity to the closest regional head (proxy for work) will hold great importance when analysing adults group accessibility. Table 4.2: Frequency of visits per month and per group to each essential facility for case study. | Category
Facility | | Youngsters | Adults | Elderly | |--------------------------------------------------|-----------------------------------------------|------------|--------|---------| | Education
Secondary schools | | 20 | 0 | 0 | | | Bachillerato schools1 | 0 | 0 | 0 | | | Universities1 | 0 | 0 | 0 | | Commerce and
Supermarkets
catering | | 0 | 4 | 4 | | Regional heads
Working
(Cabecera comarcal) | | 0 | 20 | 0 | | Healthcare | Primary medical
attention centres
(CAP) | 0.38 | 0.46 | 0.99 | | | Hospitals | 0.09 | | 0.5 | | | Municipal sports
centers | 12 | 8 | 4 | | Leisure | Shopping malls | 4 | 4 | 1 | ![](_page_70_Figure_3.jpeg) Figure 4.3: Distribution of the monthly essential trips and total number of trips per group based on the established parameters. ### 4.2.4 Definition of other parameters Acceptable travel time As indicated in the Section [2.1.1,](#page-22-2) the threshold of 30 minutes is established for all the facilities. Population data The data is taken from the available census data, which provides agedisaggregated information in five-year intervals at the municipal level for 2024 (see [INE Census](https://ine.es/jaxiT3/Tabla.htm?t=68535) [2024\)](https://ine.es/jaxiT3/Tabla.htm?t=68535). Note that the groups were created accordingly. Computation of the travel times For computing the travel times, GTFS data has been used. In particular, the GTFS of the Intermunicipal buses [\(buses interurbanos\)](https://nap.transportes.gob.es/Files/Detail/1163) offered by la Generalitat de Catalu˜na. The objective of these buses is to connect municipalities among them, which seems appropriate for the scope of this study. This choice has implications that are discussed in Section [4.2.5.](#page-72-0) Importantly, this model only considers public transport, namely buses, not considering cars. The decision not to consider car is based in the view presented in Ministerio de Transportes y Movilidad Sostenible [\(2021\)](#page-114-2): "Mobility should guarantee that citizens can cover their needs, without requiring their own vehicle" (p.62). As for the bus, this choice was based in the intuition that rural areas would only dispose of bus options, not having train stations, and aiming to simplify the model. However, train services should also have been included, for more realistic results. This inclusion would require an additional step to link bus and train stops, that would be at a close distance but not share the same stop code. As for the travel times, they have been computed following the specifications explained in Section [3.2.3,](#page-55-0) but with several simplifications, aiming to reduce the computation time. This simplifications are only included due to the high computation time of the code, and are only acceptable because the objective of the thesis is rather to show the potential results that could be obtained with it, than to offer precise results for Tarragona. However, if a real study was to be carried, it would be optimal to remove them, and compute the full procedure, without shortcuts. The applied simplifications are - If direct routes were found between two municipalities, indirect routes were not computed, assuming implicitly that direct trips will always be faster than those with transfers. - When a direct route was found between two regions, only one origin destination pair was calculated. This means that if a route has several stops available at the start or end municipality, only one pair is calculated, assuming that the main component of the travel time is the time it takes to connect these municipalities and not the travel time inside it. - Once an indirect route was found between two municipalities, the program stopped computing other combinations for that pair of municipalities. This simplification makes that the travel time for indirect routes may not be the actual minimum, and that the values should rather be understood as maximum values. Therefore, only in those cases were no solution was found, all pairs were checked. Apart from that, two additional considerations were also taken into account. • If the travel time to any essential facility was higher than the travel time to hospitals or to the regional head, it was checked if the regional head or the hospital municipalities had the essential facility under study. If so, the travel time was modified accordingly, selecting the shortest one. This is a correction that is necessary, specially for those essential services that are in numerous locations. In the algorithms, only the three closest municipalities with the service are analysed, so there is no guarantee of selecting the regional head or the hospital municipality (which are usually more important or central). However, it may be the case that the regional head may offer better connection than the surrounding municipalities (this is, it may be easier to reach the regional head by public transport than the neighbouring municipalities). This mechanism check this, ensuring logical and consistent results. • An additional consideration was also imposed: a maximum public transport travel time of 3 hours. Trips exceeding this threshold were considered infeasible and excluded from the analysis. The implications of these simplifications are explained in the next Section. ### 4.2.5 Implications: a public transport indicator for rural areas As discussed above, several choices need to be made in order to compute the indicator, and these choices have implications. In this work, two main choices make that this indicator is more suited for rural than for urban areas, and focuses in public transport. The first one is the choice of the aggregation level per municipality, which is suited to analyse rural areas, but is not precise enough to compute travel times within urban areas. This is mainly due to the computation method, explained in [3.2.3.](#page-55-0) Since this method allows to compute the travel times between two areas, but not within each of these areas, if the municipality is chosen as the level of aggregation, it is possible to compute travel times between municipalities but not within them. This choice also implies that it is considered that all the stops in a municipality are available for all its residents, and similarly that all stops in a municipality of destination provide access to all its services. The first part of this assumption is more likely to hold in rural areas, where either the bus lines usually cross the whole municipality, providing easier access, or people are willing to find ways to get to the stop they need (for example asking for someone else to drive that person to the bus stop). However, the assumption is less strong in urban areas, where the proximity to numerous bus stops and the presence of dense traffic discourages people from travelling long distances within the municipality to access transport stops. Further discussion on how to overcome this limitation is presented in Section [5.2.3.](#page-94-0) Also in relation with the geographic choice, the fact of only considering the province of Tarragona limits the validity of the results. In this study, the province of Tarragona is treated as an isolated unit, meaning that only services located within the provincial boundaries are considered accessible. This creates a border or island effect, whereby municipalities near the edge of the province may appear to have limited access to services, even though they might actually rely on facilities located just outside the study area. These external connections are not captured in the analysis. To overcome this issue, the solution would be to consider a broader geographical scopes with more areas of study. Secondly, the decision to include only the GTFS data for inter-municipal buses is influenced by the will to focus on public transport, and entangles that results are best suited for rural areas. Regarding the first will, the choice to consider only public transport and not cars (which are commonly used in rural areas) is based on the Ministerio de Transportes y Movilidad Sostenible [\(2021\)](#page-114-2), which states that "mobility should guarantee that citizens can cover their needs without requiring their own vehicle" (p.62). This decision is also endorsed by the emphasis on promoting and prioritizing more sustainable alternatives. Studies such as Ritchie [\(2023\)](#page-114-3) show that the carbon footprint per kilometer traveled in 2022 is lower for public transport (e.g., bus, train, tram) compared to private cars, even electric ones. For example, while a person emits the equivalent of 27g of CO2 to travel 1 kilometer by bus, using an electric vehicle that person would emit 47g of CO2, and using a traditionally fuelled car, it would emit 170g. However, because of only including GTFS from interurban buses, excluding urban transportation means, results are more suited to rural areas. In most rural areas, interurban buses represent the only way of reaching essential destinations by public transport. However, in urban areas, other urban means of transport are available such as metro, tramway or urban buses. However, adding these means of transport would highly increase the complexity of the analysis, and difficult the data gathering process. Therefore, it is considered that because of only considering this data, the results are more appropriate for rural than for urban areas. Despite this, it is noted that others means of interurban transport such as trains should also have been included, for more realistic results even for rural areas. Therefore, the choices made imply that, while the results can be (and are) computed for both, rural and urban areas, due to the level of aggregation chosen, and of only including inter-municipal bus GTFS the results are considered to be more reliable for rural areas. In addition, the limited computing power, which led to the simplifications adopted, and the none inclusion of other GTFS data than bus, make that these results should be understood as provisional or as an example of what could be achieved, rather than as definitive and fully operable results. In addition, for the purpose of this thesis a random week day was chosen to perform the analysis: Thursday 20th of March 2025. This day represents an average working day [\(Calendar of festivities 2025\)](https://treball.gencat.cat/ca/ambits/relacions_laborals/ci/calendari_laboral/calendari-festes-2025/). Therefore, results may vary slightly if another day was taken. As a final remark, the results are highly dependent on data quality. For example, certain bus routes may exist in reality but not be included in the GTFS files used for this study. This is expectable, as there is not currently an authority in Spain confirming that the data voluntary uploaded is accurate. For instance, when analysing the GTFS data uploaded by the Catalan Generalitat authority, it was found that some buses were scheduled after 24h, or the coordinates of a few bus stops were in the middle of Africa. This undermines the reliability of this data. A similar limitation applies to the dataset of essential facilities, where some of them may be missing or outdated. # 4.3 Intermediate results: results per category This section presents the results obtained by applying the methodology described in Section [3.2.1.](#page-48-1) Note that, according to the GTFS available data, 13 municipalities in Tarragona out of 184 have no registered bus stops, and one municipality has no operating public transport services on the selected analysis day. For these 14 municipalities (7.6% of total), no information could be outputted, with the methodology exposed in Section [3.2.3,](#page-55-0) and therefore are represented in black on all maps. In the following subsections the maps are presented following the order proposed in Table [4.1.](#page-67-0) ### 4.3.1 Results for healthcare As indicated in Table [4.1,](#page-67-0) the healthcare category is composed of three types of facilities: hospitals, primary medical attention centres (CAPs), and municipal sports centres. Each of these plays an important role in supporting public health. However, their usage frequency differs significantly. In particular, citizens are encouraged—according to World Health Organization guidelines (World Health Organization, [2020\)](#page-115-10)—several times a week, far more regularly than they visit a hospital or even a doctor. Consequently, when computing overall accessibility, the weight assigned to sports centres is greater due to their higher frequency of use. Despite this, the accessibility to each of the three types of facilities provides valuable insight on its own. Figure [4.4](#page-74-0) illustrates the accessibility to hospitals across Tarragona. In total, ten municipalities host at least one hospital (indicated in white on the map). While nearly all municipalities have at least one direct public transport connection to a hospital, about 35 % of them are located more than 30 minutes away, and 13 of them do not have any direct route connecting them. Furthermore, four municipalities have no feasible public transport connection to any hospital (apart from the 14 with no public transport), highlighting important accessibility gaps in the province's healthcare infrastructure. ![](_page_74_Figure_3.jpeg) Figure 4.4: Accessibility by bus to hospitals. The situation is notably different when it comes to primary medical attention centres (CAPs). Most municipalities either have their own CAP or are located within 30 minutes of one by public transport. However, 27 municipalities—(15% of the total), experience longer than 30 minutes travel times (see Figure [4.5\)](#page-75-0). Only two municipalities do not have a direct route connecting them to a CAP, and 3 of them can not reach a CAP in less than 180 minutes. A notable pattern is that most of these less-accessible areas are situated near the provincial borders. This may be a consequence of the island approach adopted in this study, where only services within the province of Tarragona are considered. In practice, residents of these municipalities might access healthcare facilities located just outside the provincial boundaries, which are not included in this analysis. Despite hospitals and CAPs having a lower weight in the overall accessibility index due to their lower visitation frequency, ensuring timely access to both CAPs and hospitals remains crucial from a public health perspective. Providing access to these essential health services within 30 minutes should be a priority, regardless of how frequently they are used in comparison to other services. This suggests that, maybe not only the overall accessibility results should be looked up in the study, but that also some sub results may also be of interest. For instance, looking at Figure [4.4,](#page-74-0) it seems that it could be a good idea to open a hospital in the are of l'Ametlla del mar (south) to serve the surrounding municipalities which are in red despite already having direct routes to hospitals. ![](_page_75_Figure_2.jpeg) Figure 4.5: Accessibility by bus to primary medical attention centres (CAPs). Finally, considering access to municipal sports centres, Figure [4.6](#page-75-1) show that results are overall positive. All municipalities in Tarragona—except for L'Argentera (municipal code 43017) have access to a sports centre within 30 minutes by public transport. Given that sports centres are visited more frequently than other health-related facilities (as recommended by the WHO, see World Health Organization, [2020\)](#page-115-10), this widespread accessibility will have a strong positive influence on the overall accessibility score for almost all municipalities. Note that in this case, some of the municipalities with no public transport do have sports centres, so they can access them, and therefore are categorized in green. ![](_page_75_Figure_7.jpeg) Figure 4.6: Accessibility by bus to municipal sports centres. ## 4.3.2 Results for work In this analysis, employment opportunities have been proxied by the regional heads of the comarcas which typically concentrate jobs and serve as administrative and economic centres within each region. This approach corresponds to the second proposed option for facilities that do not concentrate the provision of a service (see Section [3.2.3\)](#page-55-0). As shown in Figure [4.7,](#page-76-2) there are 10 such municipalities, although they do not exactly overlap with those hosting hospitals. 122 municipalities (66%) have access to their corresponding regional head in less than 30 minutes by public transport. However, 48 municipalities still lack such access—some of them even bordering a regional head—highlighting gaps in regional connectivity despite their geographic proximity. ![](_page_76_Figure_4.jpeg) Figure 4.7: Accessibility by bus to regional heads (Cabezas comarcales), a proxy for work. # 4.3.3 Results for education Given the selected persona chosen to represent this target group (a 15-year-old student) this analysis focuses on access to secondary education. While access to Bachillerato and University is not analysed in depth here, the related calculations have been made and are accessible in the submission documentation. As shown in Figure [4.8,](#page-77-2) 47 municipalities in Tarragona have at least one secondary school, resulting in broad coverage across the province. Most areas appear accessible (shown in green), and all of them show direct access to secondary school yet some zones lack nearby institutions, potentially warranting the establishment of new schools if the local population justifies it. In addition, there are several municipalities situated close to secondary schools but still lacking adequate public transport connections. In these cases, improved transport services could bridge the gap. It is also possible that some schools are already served by private or school-specific transport, but that these are not included in the GTFS dataset. In total, 19 municipalities—along with 14 others lacking any public transport—require attention to ensure that students can reach their mandatory education within a reasonable travel time. ![](_page_77_Figure_1.jpeg) Figure 4.8: Accessibility by bus to high schools. ### 4.3.4 Results for commerce As shown in Figure [4.9,](#page-78-0) access to supermarkets is acceptable in 143 municipalities (78%). However, 28 municipalities, especially concentrated in the central northern region of Tarragona, do not have access to a supermarket. This could be partly explained by the use of an unreliable source (OSM), where too strict requirements may have been imposed, or by the fact that these municipalities may access supermarkets in another province. In any case, this category may be the least important for the following reasons: - In all working areas, there are supermarkets. Therefore, when people go to work, they can visit the supermarket, mitigating the need to have one close from their home location. However this would still be problematic for elderly that do not work (it was assumed in this study that youngsters did not go to supermarkets). - Supermarkets are large facilities that need a minimum demand to operate. Not having a supermarket does not mean lacking access to basics. Small shops may be available, even if usually they are more expensive than big shopping surfaces. - In some cases, rural areas receive special services weekly. For instance, in Conesa (municipality in Tarragona), there is a weekly service by truck of fruits and vegetables, ensuring access to basics (Llorens, [2020\)](#page-113-11). This is a commonly used configuration for rural areas. Considering this, a debate could be open to re-assess if it is required to include supermarkets in the list of essential services and facilities. ### 4.3.5 Results for leisure The results for leisure services received the lowest score. While 8 municipalities have this service, only 78 municipalities have access to them within a reasonable amount of time. The other 106 (92+14) municipalities would not be able to reach it in a reasonable amount of time. ![](_page_78_Figure_1.jpeg) Figure 4.9: Accessibility by bus to supermarkets. This may lead to inequalities in access to cultural activities such as cinemas and shops, potentially perpetuating the cultural accessibility gap between rural and urban areas. Although often seen as a dispensable service, having access to places of leisure and entertainment is important for everyone, and has a wide range of beneficial health effects such as less stress, lower heart rate, and better well-being (Zawadzki et al., [2015\)](#page-115-11) . Therefore, Tarragona should reconsider providing access to shopping malls, or the considered facilities should be expanded to include other ways of accessing culture. ![](_page_78_Figure_4.jpeg) Figure 4.10: Accessibility by bus to shopping malls. # 4.4 Final accessibility results ### 4.4.1 Accessibility general results Having computed the travel times and completed all the required columns for completing the suggested indicator in Table [3.1,](#page-51-0) it is now possible to assess the accessibility for each of the previously defined groups in each of the municipalities. Once again it is highlighted that the results displayed are a showcase of what could be obtained, and that for more operational results, a more rigorous analysis, considering surrounding regions and other means of transport (particularly train) should be adopted. Be reminded that in this section accessibility is defined and presented as the percentage of essential trips that can be done in a reasonable amount of time. Therefore 0% accessibility (in dark red in the maps), means that none of the essential trips can be done in a reasonable amount of time. Conversely 100% accessibility indicates that all essential trips can be done in a reasonable amount of time. Finally, all accessibility values below 100% imply transport poverty, as this implies that at least one essential trip can not be done in an accessible manner. However, lower percentages imply higher severities (see Section [3.2.1\)](#page-48-1). Taking for instance the municipality of Ulldemolins (code 43157), the following table, shown in Table [4.3](#page-79-2) can be obtained. This table corresponds exactly to the one proposed in Table [4.1,](#page-67-0) and is available for all the 184 municipalities, although in 14 of them only output null results as they do not have any public transport stop. Table 4.3: Complete accessibility table for Ulldemolins municipality. Examining this table allows for a deeper understanding of the methodology. The results are intuitive and easy to interpret. For instance, the first group of rows (Youngsters group) shows that essential trips for youngsters include visits to secondary schools, Primary attention centres, hospitals, sports centres, and shopping malls. In Ulldemolins, all these trips can be completed within 30 minutes, except for trips to hospitals and shopping malls (which take 68 minutes). This means that 32.4 out of the 36.4 trips made monthly by this group can be completed in a reasonable time (note that a decimal position appears if the frequency of visit is lower than once a month). Consequently, the accessibility rate is 89%, with 94 citizens in this situation, as all youngsters in Ulldemolins share the same public transport network, and therefore have associated the same travel times. In addition, as can be seen, this high accessibility is not the same for all groups. Youngsters have much higher accessibility than for instance elderly or adults. This means that, while all groups in this municipality suffer from transport poverty, adults are more severely affected by it than youngsters or elderly. Building from the results for each group, it is also possible to obtain a general accessibility indicator for Ulldemolins. In this case, if there are 94 Youngsters with an accessibility of 89%, 4042 adults with an accessibility of 23% and 511 elderly with a total accessibility of 48%, the weighted average accessibility for the municipality is of 27.2%. As can be seen, this measure reflects closely the overall needs for this municipality, as most of its citizens have very low accessibility (they can only perform 27% of their essential trips), meaning that Ulldemolins suffers from severe transport poverty. This value is available for all municipalities, and using it, Figure [4.11](#page-81-1) can be obtained. The particular relevance the frequency has in this calculation can also be investigated with this example. In this case, the only essential trips not complying with the reasonable amount of time threshold for youngsters are the ones going to shopping malls and to hospitals, so only 4.1 trips a month can not be done. However, if secondary school trips also failed to meet this requirement, accessibility would drop from 89% to 45%, as these trips are far more frequent (20 times a month). Conversely, limited access to hospitals would have minimal impact due to their low frequency. Therefore, previous results per essential service, should be considered as part of the final solution, as they provide valuable insights. Analysing Figure [4.11,](#page-81-1) it can be seen that, while there are large areas with an accessibility of 100%, mainly surrounding Tarragona (the Capital) and the region surrounding Tortosa (south-west), most of the muncipalities (116 out of 185) suffer from trasnport poverty, with several areas being severely affected. Looking again at the Figure [4.2,](#page-65-2) it is however possible to correlate that the areas that are worse served are usually those with lower population. This means that it may be useful to expand the model to account for this fact. This possibility is discussed in the next Chapter, Section [5.2.1.](#page-90-1) Another notable observation is that the municipality of L'Argentera in the center, has 0% accessibility. As illustrated in Table [4.4,](#page-82-1) this is due to the fact that, despite having available public transport that day, some services are unreachable (not possible to access by public transport in less than 3 hours), and those that are reachable take more than 30 minutes. Note, however, that this municipality has only 128 inhabitants. The same image is presented but only including rural areas, in Figure [4.12.](#page-82-0) When comparing this figure with the previous one, it is visible that most of the regions that had better accessibility are gone, since they corresponded to urban areas, so in general the map is less green. Indeed 107 of the 116 municipalities suffering from transport poverty belong to rural areas, and considering that there are 125 rural areas in Tarragona, 85.6% of the rural areas in Tarragona suffer from transport poverty. This map also helps to visually understand that it seems that the rural areas could be divided in two: those suffering a moderately severe transport poverty situation, with accessibility levels around 80% and those enduring severe transport poverty, in which 50% or more of their essential trips can not be done in an accessible manner. It is clear that areas that are very close from urban centres have much higher accessibility than those that are further away. However nuances can be seen. For instance while Passanant i Belltall is further away from an urban area than La Bisbal de Falset, the first is much better served, having higher accessibility. This is probably due to its proximity to the regional head Montblanc. It is also relevant to remember that for every municipality, not only this map is the ![](_page_81_Figure_1.jpeg) Figure 4.11: Accessibility for all municipalities. result, but also a particularized table (similar to the one provided for Ulldemolins), awarding precise information of which are the services the population of these municipalities do not have access to. This should provide local authorities with important knowledge to help them allocate resources efficiently. ### 4.4.2 Accessibility for youngsters, adults and elderly Apart from studying the overall analysis, interesting conclusions can also be drawn plotting the accessibility levels for each group. This is shown in Figure [4.13](#page-83-1) for the whole region, and in Figure [4.14](#page-83-2) focusing only on rural areas. Looking at these new plots, it can be seen that accessibility for Adults is generally lower than for other groups. This is primarily due to work-related factors, which only affect adults. Adults often need to commute to workplaces that are not as widely distributed as schools, leading to lower accessibility scores. This lack of accessibility could be mitigated by the fact that adults are likely commuting by car, which may be the cause or the consequence of having a public transport not suited for their needs. In addition, this low accessibility may be reduced if home work is increase, as the frequency of trips would diminish. Finally, the island effect of the analysis could be significant, as it is possible that some people go to other municipalities to reach their needs, and are therefore not well represented in this analysis. Table 4.4: Travel times for the municipality of L'Argentera. ![](_page_82_Figure_4.jpeg) Figure 4.12: Accessibility per municipality in rural areas. However, this results also allow for another relevant discussion. The current indicator only includes public transport. However, namely for youngsters, this is often the only way to travel independently, and therefore the results for youngsters can be presented as final results, as no other mobility options are at hand. Concerning adults, this map could largely explain why the car dependency is so high in rural areas. If these areas are not well connected to their essential destinations by public transport, people will find other ways to satisfy their needs, even if they are more polluting. More on thsi view is presented in Section [5.4.1](#page-101-1) Finally, as can be seen, removing the urban areas does not subtract substantial information, as this calculation was though and offers insightful results primary for rural areas. And this is the case, out of the 116 municipalities identified as suffering from transport poverty, 107 (more than 90%) are in rural areas. ![](_page_83_Figure_3.jpeg) Figure 4.13: Accessibility per municipality for different age groups. ![](_page_83_Figure_5.jpeg) Figure 4.14: Accessibility per municipality for different age groups in rural areas. ### 4.4.3 Policy recommendations Looking at previous Sections, if these results were taken as reflecting faithfully the situation in Tarragona, the following outputs could be obtained: • In Tarragona there are 14 municipalities with no public transport. This should be fixed as soon as possible, to ensure that the inhabitants of these areas can access the public transport network. - In Tarragona, 116 out of 184 (63 %) suffer from transport poverty only looking at accessibility, as they can not perform all their essential trips in a reasonable amount of time. iF focusing only in rural areas, 107 out of 125 municipalities (85.6%) suffer from transport poverty. Out of these 116 however, not all the regions suffer from the same poverty levels. The biggest severely affected by low accessibility area is in the northern region of Tarragona , around La Bisbal del Falset and Ulldemolins, where inhabitants can only perform between 20 and 40% of their essential monthly trips. There is also a surprising area near El Vendrell (south-east) with a low accessibility region surrounded by a high accessibility region, and a region in the south where a similar (but slightly less severe) issue occurs. This should be investigated to see if the areas are actually poorly connected, or if the data available is incorrect. The western region should also receive attention, as it also concentrate municipalities with severe transport poverty caused by low accessibility. - For 64 municipalities, the travel times to hospitals are too long. The network of buses offering access to these facilities should be re-thought. To cover distances in a "fast, reliable, comfortable and cost efficient" way (Association et al., [2004,](#page-109-13) p.1), systems as Bus Rapid Transit (BRT) could be implemented. - The regional heads, which hold great importance for the adults (and thus general) accessibility levels, but also concentrate other essential facilities, are not well enough connected. Each "comarca" should pay significant effort and redesign the public transport network to ensure that all municipalities inside a "comarca" can access its head in less than 30 minutes. - Leisure areas are not well enough connected, and are not sufficiently distributed among the territory. All of them are concentrated near the coast, so it may be beneficial to promote the cultural facilities in the northern part of the region. - In general , accessibility in rural areas is significantly lower than in urban areas, and accessibility in east Tarragona is generally greater than in the west. If the objective is to reach a 30-minutes-country, significant efforts will be required. - Particularized results at a municipal level can also be obtained. For instance, the municipality of Ulldemolins lacks access to several essential facilities, namely to Hospitals, shopping malls, supermarkets and work. All these problems would be solved at once if the municipality was better connected to its regional head. As can be seen, the results outputted are precise, comprehensive, easy to understand and actionable. Of course the solution for each problem is not easy, neither obvious, but the proposed methodology offers a clear overview of the state of the art, providing decision makers with strong and data-based arguments to prioritize investments. Therefore, the final policy recommendation would be for the decision makers to evaluate this proposal, and determine if they consider it offers a good basis for their purposes. If this is the case, then the suggestion would be to collaborate with academia and invest in developing this sort of analysis, further refining the methodology. # 4.5 Conclusions This case study demonstrates that the theoretical methodology from Chapter [3](#page-44-0) can be effectively implemented to draw conclusions. First of all, it is needed to bear in mind that in the context of transport poverty, accessibility analysis should only be performed for those means of transport previously classified as available, adequate and affordable, and accessibility should only be assessed to essential services. The criteria used concerning this matter in the study was: all stops in a municipality were considered available for its residents, and adequacy and affordability requirements were assumed to hold. As for the essential services, the list offered in Radics et al. [\(2024\)](#page-114-0) was adopted. In this context, accessibility is the last step of the transport poverty analysis, and only if it is of 100%, the citizens of that municipality can be considered as not suffering from transport poverty. As for the measurement itself, the accessibility indicator offers a unique, comprehensive measure, allowing to quickly identify areas with poor accessibility and therefore transport poverty. The proposed methodology offers a clear and precise overview of the state of the art of transport accessibility in a region, providing decision makers with strong and data-based arguments to prioritize investments. In addition, it provides useful and direct insights to understand which areas are more or less in need, and which are the concrete unmet needs for each municipality. The results also are capable of offering general trends insights, and visualize problems that may have not been identified previously. The results and insights of this study can be compared to other recent studies shading light over similar issues. When comparing these results with the foundational reference of Cludius et al. [\(2024\)](#page-109-0), it is believed that this study offers more comprehensive insights. The EU report focuses solely on travel time to work, providing no information for younger or elderly populations. Additionally, this thesis includes a colored map, quickly highlighting areas needing attention—something the EU report lacks, as it is intended for the whole EU territory and not a specific area. Similarly, the Instituto Geogr´afico Nacional (IGN) [\(2022\)](#page-112-8) report, which investigates car accessibility and distances to cities of various sizes, is less intuitive for identifying transport poverty. In this document, the proposed results allow to identify the travel times to cities having more than 5,000, 20,000 or 50,000 inhabitants. In the report, this choice is motivated by correlating that greater cities usually have more services. However, this methodology makes it difficult to infer when there is transport poverty, as for instance, not having access to a 20,000 inhabitants city does not correlate directly to suffering from transport poverty. In the proposed methodology, the focus in essential services allows for a more straightforward relationship. In this sense, if accessibility is of 80% it means that a person can perform 80% of the essential trips in a reasonable amount of time. Similarly, if a person can not perform any essential trip, then it is suffering from transport poverty (which can be more or less severe). Finally, the results are also compared to the ones in Radics et al. [\(2024\)](#page-114-0). In this case, it is also believed that the results outputted in this methodology may be more applicable. In that publication, a cumulative opportunity measure is presented (see Figure 16 of that publication), enabling to identify which areas have accessibility to more services in the same amount of time. However, this measure does not allow to differentiate if all services are available at a glance. The methodology proposed here does so, because if all essential services can be reached, accessibility will be of 100%, regardless of how many facilities of each kind can be visited in the reasonable amount of time. Therefore, it is believed that the objective for decision makers is easier in this work: to have 100% accessibility in all regions. This is less evident with the cumulative opportunity measure, where it is unclear if the desire is to have more destinations available for all areas. For example, in rural areas it would be very difficult, and maybe not desirable, to have a similar density of supermarkets or hospitals as the one present in the city centre, as there would be no demand to sustain this number of facilities. Finally, the methodology demonstrates that the choices made during its operationalization can significantly influence its alignment with European and Spanish directives. In this thesis, strategic decisions, such as including only public transport, ensure that the results promote sustainable modes of transportation, rather than private vehicles. Similarly, selecting municipalities as the level of aggregation makes the results more insightful for rural areas, emphasizing their importance in the discussion. These choices are important, as they will shape the debate, and shift focus. For instance, if cars had been included in the analysis, the results would probably had headed authorities towards building new roads or increasing the capacity of existing ones. However, when considering public transport, the debate shifts, and tends to pay more attention to matters such as how to optimize fleets and routes, or how to make vehicles move faster. # Chapter 5 # Discussion Having applied the methodology in a real case scenario, this Chapter first evaluates the strengths and limitations of the proposed methodology. After that, proposals on how to improve it are discussed. To demonstrate the replicability of the indicator, an explanation of how this methodology could be applied in another country (in this case the Netherlands) is provided. Finally, a short discussion on which other results (non directly related to transport poverty) could be obtained when applying this methodology are exposed. # 5.1 Strengths and limitations of the methodology # 5.1.1 Strengths of methodology The first identified strength of this methodology is that it offers a working definition of transport poverty, based in the current literature (and mainly Cludius et al., [2024\)](#page-109-0), that allows for the creation of a new framework to investigate transport poverty. This new framework, exposed in Figure [3.1,](#page-45-0) establishes some sequencing of the different dimensions of transport poverty, allowing to structure the different issues, and offering a systematic way to evaluate if an individual suffers from transport poverty. Since the suggested workflow is inspired on the decision making process of a person, it should be easy to follow and understand by the general public, allowing to democratize the concept of transport poverty, and enhance public debate on the matter. Finally, this sequencing could also help decision makers when prioritizing investments, offering some guidance on which matters should be addressed first. Yet, it is reminded that this framework is only proposal, and requires validation. A second strength identified in the work is the development and application of an indicator related to accessibility, adapted for the context of transport poverty. This methodology, inspired and based in previous literature (and mainly Radics et al., [2024\)](#page-114-0), offers an indicator that measures accessibility as the percentage of essential trips that can be done by citizens living in each area. This measurement therefore allows to quickly identify if there is transport poverty (accessibility is less than 100%) and its severity. In addition, it offers an easy to interpret indicator, which is not only helpful for decision makers, but also for the general public, that will receive messages and information that can be understood without requiring (too much) technical knowledge. The proposed methodology offers therefore dichotomous results (poor or not poor), but also allows for nuances when establishing the thresholds, and the results allow to differentiate for different poverty levels (severity). The choice of offering dichotomous results is considered to be suited for decision makers, that usually require direct statements and clear conclusions to decide upon investments. However, the possibility to nuance the poverty thresholds, hence influencing the final result, also offers a framework that is flexible enough to classify poverty correctly. This is mainly because the indicator requires the decision upon numerous parameters, which may be accurately adapted to classify correctly which regions are poor. Another advantage of the offered methodology is that it can be adapted for different groups and regions, and still provide comparable results. For instance, while it is clear that different age groups have different needs and require accessing different destinations (see Figure [4.3\)](#page-70-1), the proposed framework allows to compare all of them in a same scale and at a glance (see Figure [4.13\)](#page-83-1). A third identified asset of the proposed methodology is that it is capable of outputting results aligned with the directives established by the European and Spanish institutions. As detailed in Chapter [4,](#page-62-0) this work makes the choice of analysing public transport accessibility, as opposed to other recent studies such as Instituto Geogr´afico Nacional (IGN) [\(2022\)](#page-112-8) that focus in car accessibility. Therefore, results align more closely with the European Green Deal ambitions (European Parliament and Council, [2019\)](#page-111-3), that promote the development and prioritization of the most sustainable ways of transport, to accelerate decarbonization. This choice is also aligned with the Spanish mobility strategy (Ministerio de Transportes y Movilidad Sostenible, [2021\)](#page-114-2). In addition, the decisions on the aggregation level of the case studied and the networks of transport included, also make this indicator more reliable/suited for rural areas (Section [4.2.5\)](#page-72-0). Hence emphasising their needs, and therefore paying attention to vulnerable groups in this context (European Parliament and Council, [2022\)](#page-111-7). Finally, this work also promotes the use of open data, showcasing how different sources of information can be merged and integrated to output unique and insightful results. ### 5.1.2 Limitations of methodology Despite the notable strengths of the proposed methodology, several limitations must be acknowledged across different dimensions. First, further research and consensus are required regarding the conceptual foundations of transport poverty. Key components—such as the definition of essential services and essential trips, the dimensions of transport poverty, and appropriate thresholds for what constitutes a "reasonable" travel times, are still subject to debate. For example, a recent study on transport poverty in outermost regions (Maucorps et al., [2025\)](#page-113-12) introduces additional dimensions, including time poverty or externalities of transport, which were not addressed in this work. Similarly, while the definition of "essential services" is an actively discussed topic in the literature, there is no universally accepted standard. These conceptual ambiguities highlight the need for continued academic and policy dialogue to establish well-founded, broadly accepted definitions and thresholds that enhance the reliability and comparability of accessibility indicators. Another significant constraint relates to data availability and quality. Although the use of GTFS (General Transit Feed Specification) data has become increasingly widespread, coverage remains inconsistent across regions. As of May 2025, several Spanish autonomous communities—including Murcia, La Rioja, Castilla y Le´on, and the Canary Islands—do not provide data for inter-municipal bus services. Moreover, GTFS datasets often contain inaccuracies or incomplete entries, such as misplaced stops or outdated routes, which undermine their reliability and may not accurately reflect current public transport networks. Regarding essential facilities (providing access to essential services), while efforts have been made to consolidate location data (for example ,the Spanish government has developed a national database of facility locations, used in this study, known as the Points of Interest Database, see: [IGN BTN-POI\)](https://centrodedescargas.cnig.es/CentroDescargas/btn-poi) official datasets are still not universally available or comprehensive. Besides that, the greatest data gap remains in understanding how frequently individuals need to make use of essential services. Information on trip frequency is vital for aligning public transport provision with actual needs, yet such data is rarely collected systematically. To improve transport planning, the collection and publication of data on citizens' service usage patterns should become a standard practice. A further limitation pertains to the computational demands of the methodology. The implementation developed for this thesis is computationally intensive for two main reasons. First, the coding was carried out without advanced expertise in Python optimization, resulting in a code that could most likely be streamlined. Second, the nature of the algorithm itself—evaluating all origin-destination pairs across multiple potential routes, including transfers—supposes a substantial number of checks and comparisons. Efforts should be put in perfecting the code, to obtain reliable results in shorter times. Considering the indicator itself, further refinement is also needed in the treatment of work and education trips, which constitute a substantial share—approximately 50%—of all essential trips for adults and youth, as illustrated in Figure [4.3.](#page-70-1) Ideally, employment accessibility should be modelled using supply and demand dynamics rather than assuming a fixed work location. A gravity model, such as the one used by Allen and Farber [\(2019\)](#page-109-4), could be implemented to better match workers with employment centres. In this sense the Spanish National Statistics Institute (INE) provides relevant data on both job demand and supply [\(New monthly contracts by municipality,](https://datos.gob.es/es/catalogo/ea0021425-contratos-por-municipios) [Job seekers by municipality\)](https://datos.gob.es/es/catalogo/ea0021425-demandantes-de-empleo-por-municipios), which could support the development of such a model (see Section [2.2.2](#page-34-0) for more details). Similarly, in the education domain, students in Spain are typically assigned to specific institutions rather than attending the closest one. Thus, a more precise matching between students and educational centres could further enhance the accuracy of accessibility calculations. Another important limitation of the accessibility indicator is that it does not put emphasis in areas where more individuals are suffering from transport poverty, something that should be beard in mind when allocating investments. This is, the amount of people affected by transport poverty is not a criteria used to prioritize investments. However, this caveat is solvable, and it is discussed how to do it in Section [5.2.1.](#page-90-1) Similarly, the proposed methodology could also be refined, to filter better which are the concrete stops providing access to a facility, matter discussed in the following Section [5.2.3.](#page-94-0) The current model also assumes that all essential trips begin at the place of residence (see Section [3.2.3.1,](#page-55-1) thereby omitting trip chaining and real-life commuting behaviour. In practice, individuals often combine multiple activities in a single journey—for example, stopping at a grocery store on the way home from work. Capturing this level of behavioural complexity would require origin-destination data that considers multiple points of interest throughout the day, which is beyond the scope of the current model but remains an area for future improvement. Lastly, the model presumes that individuals make decisions based solely on rational, practical factors such as proximity and efficiency. In reality, personal values, social norms, and cultural preferences also play a critical role in shaping mobility behaviour (Handy et al., [2005,](#page-112-1) Van Acker et al., [2010\)](#page-115-0). This raises the complex question of whether transport poverty should be defined solely by measurable limitations in access, or also account for personal preferences and choices. In this regard, one of the unresolved ethical considerations is whether individuals who have intentionally moved to low-accessibility areas should be considered transport poor in the same way as those constrained by socio-economic limitations. This issue, while beyond the scope of this thesis, is paramount to the broader discussion on how transport poverty is defined and operationalized in policy-making. It suggests that indicators of transport poverty should incorporate not only structural accessibility but also dimensions of agency and choice, which are not addressed in the current methodology. # 5.2 Further improvements of the methodology ## 5.2.1 Expanding the indicator including number of citizens Looking at the results obtained in the previous Chapter and shown in Figure [4.11,](#page-81-1) the accessibility for each of the region can be assessed. Yet, if looking also at Figure [4.2](#page-65-2) it can be seen that most of the areas with lower accessibility levels are also those with less residents. Therefore, a clear intuition can be established: it may not be preferable to invest in a municipality with very low accessibility if the number of residents is also very low, unless the goal is to attract people to that area. To gain insights in this regard, two new indicators have been created: • Unfeasible trips indicator: The original accessibility indicator is calculated as the ratio of essential trips that can be completed within a reasonable time to the total number of essential trips. This means that, with the available data it is possible to compute the number of non complying trips (or non accessible trips), FNCim, for each group i for each area of study (in this case municipality m). This, as shown in Equation [5.1,](#page-90-2) can be obtained by subtracting to the total number of trips each group i needs to do to all L essential facilities, the number of complying trips, computed using Cijm, a binary variable that takes the value of 1 if the travel time to reach each essential facility j from each municipality m for each group i (tijm) is below the acceptable threshold to reach each facility for each group (Tij ). Note the notation used is the same as in Section [3.2.1.](#page-48-1) $$F_{NCim} = \sum_{j=1}^{L} F_{ij} - \sum_{j=1}^{L} F_{ij} \cdot C_{ijm}$$ (5.1) Then, multiplying the number of non complying trips (FNCim) by the population of each group and each municipality Pim, and summing the result for all K groups in that municipality, the total number of essential trips that remain unfeasible for all residents in that area can be obtained. This could be expressed as follows in Equation [5.2,](#page-90-3) where FNCm represents the number of non accessible trips in municipality m, FNCim represents the number of non accessible trips for group i in municipality m (there are K groups), and Pi,m is the number of residents of group i in municipality m: $$F_{NCm} = \sum_{i=1}^{K} F_{NCim} \cdot P_{im}.$$ (5.2) This measure is particularly useful because it incorporates both accessibility limitations and the scale of the affected population. For example, a municipality with 1,000 residents unable to complete one essential trip each (so in total 1,000 trips can not be done in the municipality), may require more urgent attention than a municipality of 50 residents unable to complete 10 trips each (in total 500 trips). The results of computing this indicator for the case study presented in Chapter [4](#page-62-0) is shown in Figure [5.1.](#page-92-0) It seems a relevnt indicator that could be used by decision makers to prioritize investments. • Excessive time spending indicator: Another useful indicator focuses on the amount of time individuals spend above the acceptable travel threshold when using public transport. This helps differentiate between cases where a destination is slightly beyond the threshold (e.g., 35 minutes, 5 minutes over the limit) versus significantly distant (e.g., 60 minutes, 30 minutes over the limit). By calculating the number of minutes exceeding the threshold and multiplying it by the monthly frequency of each essential trip, the total amount of excess time a person spends due to inefficient transit can be calculated. This calculation can be done with Equation [5.3,](#page-91-0) in which Hm is the total number of hours that people from municipality m spend in public transport, tijm is the time in hours it takes for citizens of group i living in municipality m, to reach the essential service j (there are L essential services and K groups), Tij is the maximum amount of time in hours it should take for citizens of group i to reach the essential service j, Cjim is a binary variable that takes the value of 1 if the travel time (tijm) is below the acceptable threshold (Tij ) and finally Pim is the number of residents of group i in municipality m. $$H_m = \sum_{i=1}^{K} P_{im} \sum_{j=1}^{L} (t_{ijm} - T_{ij}) \cdot (1 - C_{ijm}). \tag{5.3}$$ A downside of this indicator, is can only be applied to municipalities where all destinations are reachable in the first place. This is because even if only one service is not accessible (not possible to reach using public transport), the indicator can not be computed, as it would require to input data, and therefore distort the results. It is also noted that both indicators can also be normalized by the total population of each area of study to obtain per capita values, as shown in Equations [5.4](#page-91-1) and [5.5](#page-91-2) . This allows to estimate, for each municipality, the average number of essential trips per capita that cannot be completed in a reasonable time (FNCp.c.m), as well as the average excessive amount of time (in hours) each resident spends on public transport (Hp.c.m). These individual-level results provide a more equitable basis for comparison across municipalities of different sizes, not making differences across groups. $$F_{NCp.c.m} = \frac{\sum_{i=1}^{K} F_{NCim} \cdot P_{im}}{\sum_{i=1}^{K} P_{im}}.$$ (5.4) $$H_{p.c.m} = \frac{\sum_{i=1}^{K} P_{im} \sum_{j=1}^{L} (t_{ijm} - T_{ij}) \cdot (1 - C_{ijm})}{\sum_{i=1}^{K} P_{im}}.$$ (5.5) Under the light of these new indicators, the urgency of intervention changes. Focusing for instance in the aggregated results of Figure [5.1a](#page-92-0) the municipalities that need more attention are: Vandell`os i l'Hospitalet de l'Infant (43162), Ulldecona (43156), Cunit (43051), L'Ametlla de Mar (43013), and La S´enia (43044). However, as shown in Figure [4.11,](#page-81-1) the accessibility for these municipalities ranges between 40% and 60%, so they would not be the priority if only looking at accessibility requirements. A similar perception can be obtained looking at Figure [5.1b](#page-92-0) showing the results per capita. As can be observed, even though the number of trips that cannot be completed per person is not the highest of the entire province, these areas are more populated than those with lower accessibility levels (Figure [4.2\)](#page-65-2). Consequently, more people are affected by this poverty condition, leading to a higher overall number of trips that cannot be completed. When focusing solely on rural areas (Figure [5.2\)](#page-93-1), the results are very similar. This is expected, as the chosen method is more suited for rural than for urban areas,(Section [3.2.3.2\)](#page-57-0). Note that, in the real life, these results may not be realistic, as they do not take into account the train line R16 which connects most of the municipalities in the coast of Tarragona, and has stops in L'Ametlla de Mar, l'Hospitalet de l'Infant or Ulldecona (see Figure [5.3\)](#page-94-1). ![](_page_92_Figure_3.jpeg) Figure 5.1: Comparison of unfeasible trips per municipality and per citizen. onable amount of time per citizen. onable amount of time per municipality. ![](_page_93_Figure_1.jpeg) ![](_page_93_Figure_2.jpeg) - (a) Number of trips that can not be done in a reasonable amount of time per municipality in rural areas. - (b) Number of trips that can not be done in a reasonable amount of time per citizen in rural areas. Figure 5.2: Comparison of unfeasible trips per municipality and per citizen in rural areas. Figure [5.4b](#page-94-2) provides further nuance, showing that municipalities like Cunit (located at the very east) may not have high accessibility, but the travel times are only slightly above the threshold. As can be seen citizens in Cunit spend less than 5 excess hours per month in public transport, whereas in L'Ametlla de Mar (triangular shape), this time is 12.5 hours per month. This allows for new prioritization of investments, depending on the aspects considered. Unfortunately, numerous municipalities do not have access to at least one service, and therefore the indicator cannot be computed for these cases. ### 5.2.2 Expanding the indicator modifying compliance requirements Hospitals provide a useful case to demonstrate how the accessibility indicator can be further extended by incorporating two additional dimensions when computing the compliance parameter (Ci,j,m): arrival time constraints and minimum service frequency. These additions allow for a more nuanced understanding of accessibility, particularly in the context of transport poverty and system resilience. First, certain medical services, such as blood tests or specialist appointments, often require early morning arrivals. To explore this, an analysis was conducted to determine which municipalities could reach their nearest hospital before 8:30 AM within a reasonable travel time. As shown in Figure [5.5,](#page-95-1) and compared to the general accessibility map (Figure [4.4\)](#page-74-0), the number of underserved municipalities increases from 4 to 37 when this time constraint is applied. This highlights that accessibility as defined in this study does not guarantee usability under real-life time conditions. Second, the role of service frequency was evaluated by requiring at least two daily expeditions to the nearest hospital to label trips as accessible. As depicted in Figure [5.6,](#page-96-0) this threshold leads to 54 municipalities lacking adequate coverage. This is particularly relevant, for instance in cases of schedule disruptions, or to ensure that if citizens have some obligation ![](_page_94_Figure_1.jpeg) Figure 5.3: Line R16 of Rodalies going through Tarragona [[Rodalies de Catalunya](https://www.trenscat.com/renfe/r16_ct.html)]. ![](_page_94_Figure_3.jpeg) Figure 5.4: Excess time spent in public transport per municipality and per person. at one time on the day, they will still have an alternative to make their trip. These findings suggest that definitions of accessibility and transport poverty—such as those proposed by the European Commission in Cludius et al. [\(2024\)](#page-109-0)— could be extended to consider not only spatial reach but also temporal availability and number of expeditions. However, including additional requirements, could also open a debate on what is the strict minimum to consider that accessibility is sufficient, and what constitutes a desirable improvement for building a more robust, comfortable and resilient transport network. ### 5.2.3 Expanding the methodology using catchment areas for each service Another identified limitation of the methodology (Section [5.1.2](#page-88-0) is its lack of precision when considering which stops provide access to essential facilities . In Chapter [4,](#page-62-0) the aggregation level was the municipality, consequently assuming all stops within a municipality provide access to all facilities there (Section [3.2.3.2\)](#page-57-0). However, this assumption may not hold in large ![](_page_95_Figure_1.jpeg) Figure 5.5: Accessibility by bus to hospitals arriving before 8:30 AM. urban areas. Therefore, refining the catchment area of facilities using precise coordinates could yield more accurate results. To achieve this, a small adjustment in the calculation method should be done. Instead of considering a unique list of polygons that can act indistinctly as origin and departures, two different lists should be considered. The first one would be for origins, and would include municipalities or any other partition of the territory. The second would be a list of catchment areas, where each facility would be assigned a given distance (d) that would provide access to that facility. Only stops within that catchment area would be considered as providing access to the facility. This idea is illustrated in Figure [5.7.](#page-96-1) However, this method has a significant limitation. In large cities, it is common for other municipalities to lack direct lines to services, requiring users to access a central point in the city and then navigate its urban network. For example, a rural resident might arrive at the central bus station of a municipality and then use the metro or urban bus to reach a hospital. Therefore, if trips are evaluated based solely on their arrival at the exact required stop(s) without including the GTFS data of urban networks, many trips might be incorrectly deemed impossible. Additionally, including this data would increase the number of transfers, thereby raising the computational time and complexity of the process. In any case, a significant advantage of including this proposal is that it would also allow for precise analysis of the urban areas, unlike the current approach. ### 5.2.4 Other improvements To enhance the methodology, several improvements can be considered: • Include more means of transport: Incorporating additional modes of transport such as trains, trams, and other public transit options is necessary to provide a more comprehensive analysis of accessibility. ![](_page_96_Figure_1.jpeg) Figure 5.6: Accessibility by bus to hospitals only considering municipalities with more than one expedition per day. ![](_page_96_Figure_3.jpeg) Figure 5.7: Illustration of the methodology for computing travel times accounting for catchment areas. - Including a return trip requirement: Similarly to the extensions done in Section [5.2.2,](#page-93-0) another interesting addition would be to verify if it is possible to compute a return trip to the destination, instead of only checking one way. - Analysis with 1km squared grids: Performing the analysis using 1km squared grids rather than Local Administrative Units (LAUs) could allow to achieve finer spatial resolution (Cattaneo et al., [2021\)](#page-109-3) and increase comparability among the whole European union, being consistent with the Eurostat [\(2021\)](#page-111-12) criteria. • Expand the model to include other groups: Investigating how the model could be expanded to include other groups of interest, such as caregivers, would also be an interesting development line # 5.3 Applicability in other contexts, a case study of the Netherlands The methodology developed in this thesis is designed to be adaptable to various territorial contexts, provided that relevant data sources are available and certain adjustments are made to reflect local conditions. This section discusses the potential application of the methodology in the Netherlands and outlines the necessary adaptations, while also providing a critical comparison with the Spanish context. ## 5.3.1 Administrative units and population disaggregation For application in the Netherlands, municipalities should also serve as the unit of analysis. With 342 municipalities, they represent the lower level of governance with substantial administrative capacity and autonomy, making them suitable for targeted policy measures. Provinces, by contrast, are too extensive to capture the granularity required for detailed accessibility assessments. Lowest levels could also be applied if there is enough data. Population data should be disaggregated by 10-year age groups, as provided by the Dutch Central Bureau of Statistics (CBS) in the regional population dashboards [\(CBS Dashboard\)](https://www.cbs.nl/nl-nl/visualisaties/dashboard-bevolking/regionaal/inwoners). This granularity allows for an estimation of age-specific mobility needs and service demand, but allows for less disaggregation than for the Spanish case, where 5-years age groups were available. ### 5.3.2 Essential facilities selection and setting of essential trips frequencies The classification of essential facilities (dependent on the decided list of essential services) should be validated with local stakeholders to ensure it reflects the actual needs and behaviours of Dutch residents, as proposed by Logan and Guikema [\(2020\)](#page-113-7). The general categories defined in this methodology—education, work, healthcare, shopping, and leisure—remain relevant, but should be specified to suit the Dutch context. Education In the Netherlands, compulsory education extends to the age of 18, with three primary educational pathways: VMBO (Voorbereidend Middelbaar Beroepsonderwijs – Pre-vocational Secondary Education), which concludes at age 16 and often leads to MBO (Middelbaar Beroepsonderwijs – Secondary Vocational Education). HAVO (Hoger Algemeen Voortgezet Onderwijs – Senior General Secondary Education), continuing until age 17 and leading to HBO (Hoger Beroepsonderwijs – Universities of Applied Sciences). VWO (Voorbereidend Wetenschappelijk Onderwijs – Pre-university Education), which finishes at 18 and provides access to traditional research universities. Secondary schools are typically located within the same municipality as the student's residence [\(Secondary Schools\)](https://allecijfers.nl/middelbare-scholen-overzicht/). However, higher education institutions require careful differentiation between MBO, HBO, and traditional universities due to their varied spatial distribution and governance structures [\(Universities,](https://allecijfers.nl/universiteiten/) [HBO,](https://allecijfers.nl/hbo/) [MBO\)](https://allecijfers.nl/mbo/). This would make it more complex to select a representative persona for all of them, and maybe subgroups within students should be created. As for frequency parameter, the same frequency of 5 trips per week could be used. Healthcare Healthcare accessibility includes both general practitioners (GPs) and hospitals. Every Dutch municipality hosts at least one GP, ensuring a basic level of coverage across the country [\(GP Coverage Map\)](https://www.vzinfo.nl/eerstelijnszorg/regionaal/huisartsenzorg). However, hospitals are more sparsely distributed and require geolocation from official datasets [\(Hospital Locations\)](https://www.vzinfo.nl/ziekenhuiszorg/regionaal/locaties). Dutch citizens visit their GP approximately four times per year on average [\(CBS GP](https://www.cbs.nl/nl-nl/nieuws/2013/27/ongeveer-drie-kwart-bezoekt-jaarlijks-huisarts-en-tandarts) [Statistics\)](https://www.cbs.nl/nl-nl/nieuws/2013/27/ongeveer-drie-kwart-bezoekt-jaarlijks-huisarts-en-tandarts), and hospital visit frequency can be estimated using national healthcare utilization data [\(Hospital Usage Data\)](https://www.vzinfo.nl/ziekenhuiszorg/gebruik). Finally, as for sports centre, the [Database SportAanbod \(DSA\)](https://www.mulierinstituut.nl/programmas-aanbod/database-sportaanbod-dsa/) from Mulier Instituute hosts a list with 22,000 entries with detailed information from sport facilities such as facility types, locations, and the municipalities they belong to. The same frequencies used in Table [4.2](#page-70-2) can be used, as these values were based in the recommendations of the World health organization. Work and Economic Activity To identify key employment centres, choosing municipalities within the Randstad area and province capitals could be a good option. Alternatively, a threshold can be applied—such as municipalities with over 100,000 jobs—to define work destinations in the following source: [\(Employment Statistics by Municipality\)](https://www.allesupermarkten.com/plaatsen/). The same frequency as in Spain could be used. Shopping and Leisure Grocery accessibility can be assessed using data from OpenStreet-Map or commercial aggregators, such as [Allesupermarkten,](https://www.allesupermarkten.com/plaatsen/) which provide comprehensive lists of supermarket locations. Leisure and shopping mall accessibility can be determined using commercial location databases, such as [CommasCom,](https://www.commascom.nl/content-cijfers.html) or again using OpenStreetMap. In both cases, the same frequencies as in Spain could be used . ### 5.3.3 Time thresholds adapted for a more urban context Unlike Spain, where rural and intermediate regions are more prevalent, the Netherlands is predominantly urbanized. As illustrated in Figure [5.8,](#page-99-1) the EU's urban-rural typology classifies most Dutch municipalities as predominantly urban (Eurostat, [2025\)](#page-111-0). However, this classification is not universally accepted. For instance, the Netherlands has is own classification considering the degree of urbanization, as shown in Figure [5.9,](#page-99-2) showing that there are rural areas in the Netherlands. This evidences the relevance of Section [2.1.3,](#page-27-0) that highlights that the definition of rural areas should be defined for each case. Despite this, it is clear that the Netherlands is a small and predominantly urbanized country, and therefore the threshold for a reasonable travel time can be adjusted to reflect the higher population density, the shorter distances and a more robust transport network. Consequently, a maximum travel time of 15 minutes is proposed, aligning with the 15 minutes city theory (Pozoukidou and Chatziyiannaki, [2021\)](#page-114-13). ![](_page_99_Picture_2.jpeg) Figure 5.8: Urban-rural typology according to EU classification. Predominantly urban regions (blue), intermediate regions (brown), rural areas (green). [Eurostat, [2025\]](#page-111-0). ![](_page_99_Picture_4.jpeg) Figure 5.9: Classification of the degree of urbanization according to the [CBS classification.](https://www.cbs.nl/en-gb/onze-diensten/methods/definitions/degree-of-urbanisation) Red: extremely urbanised, blue: strongly urbanised, light green: moderately urbanised, dark green: hardly urbanised , pink: not urbanized (equivalent to rural areas). ### 5.3.4 Data availability and model transferability The high level of open data availability in the Netherlands greatly facilitates the transferability of this methodology. GTFS feeds are available for most public transport operators, and demographic and spatial data are well-documented and publicly accessible through platforms such as [CBS](https://www.cbs.nl/) and [PDOK](https://www.pdok.nl/) (Publicly available geodata). In summary, the methodology proposed in this thesis is applicable to the Netherlands, provided that suitable modifications are made to accommodate national essential services. The high data availability should make this methodology easily applicable, and allow for a greater level of detail than in the Spanish case. However, the model should encompass as much data as possible, and not only buses, as in the Netherlands most of the territory is classified as an urban area. In addition, stricter requirements to stations could be applied (for instance only consider as possible destinations those stops within 500 m from a service), and cycling should be considered when performing availability analysis ### 5.3.5 Transport poverty in the Netherlands, a relevant issue Transport poverty is a significant issue in the Netherlands. In 2018, the Central Bureau of Statistics (CBS) introduced an indicator to measure it (Kampert et al., [2018\)](#page-112-13). Following this, on April 23, 2019, the House of Representatives passed a motion requiring the government to consider transport poverty and the CBS indicator in public transport policy. The CBS then refined its method and published a second report in 2019 (Kampert et al., [2019\)](#page-112-14). The CBS methodology aims to consider transport poverty comprehensively, including:(1) Means of transport (availability), (2) Proximity to destinations (accessibility, the focus of this thesis), (3) Income (relating to affordability) and (4) Household characteristics (which relate to adequacy). These components align with those proposed in the foundational EU Commission report (Cludius et al., [2024\)](#page-109-0). Each category has two or three indicators measured at the household level (9 in total), that take categorical values from 0 (no contribution to transport poverty) to 2 (high contribution to transport poverty). The the average level is computed and used to determine the risk of transport poverty. Regarding accessibility indicators, the CBS methodology emphasizes the importance of proximity to work, education, care, shops, and social contacts for societal participation. The first four elements are also present in the current methodology, but the fifth is not, as the proposed model does not include social contacts as an essential destination. Including relatives households as an essential destinations could indeed be very appropriate, and makes this approach more complete than the current one. Yet it would be very difficult to replicate this in other areas, as it is rare to have precise information on the location of the population relatives at a major scale. In contrast, an advantage of the proposed methodology in this document versus the proposal of the CBS is that it is capable of assigning different facilities to different groups, not having to select generic ones that fit all of them. In addition, work category is also added (even if proxied) in the methodology, but excluded in the CBS report. A major change in the proposed methodology in this thesis compared to the CBS report is the sequencing of the dimensions of transport poverty. In Figure [3.1,](#page-45-0) it can be seen that transport poverty dimensions are sequenced to provide order when addressing different issues. For example, according to the proposal, it would not make sense to compute accessibility by public transport in areas where it is not available or affordable. First, those issues need to be resolved, and then accessibility analysis can begin. The CBS model computes everything at once. This makes the results more informative, as they provide insights on all dimensions, but they also make results harder to interpret and act upon. This is confirmed looking at the discussion section, that highlights that "it is complex to interpret the outcomes of the indicator" (Kampert et al., [2019\)](#page-112-14). For instance, if a household has low income and low accessibility to hospitals, the CBS model does not link these concepts, making it difficult for decision makers to know what to do about it and where does the problem lie exactly. Despite this, the CBS report provides valuable information, highlighting the importance and worthiness of investigating transport poverty and the need for open data to achieve meaningful results. Other reports such as voor de Leefomgeving [\(2022\)](#page-115-12) also allow relevant insides for accessibility in this country. # 5.4 Other results of interest In Chapter [4](#page-62-0) it has been proved that this indicator can provide insights in accessibility. This Section examines if other results of interest could be obtained using the exposed methodology. ### 5.4.1 Comparison with car-based accessibility measures Another relevant result of this study is to compare how accessibility by public transport compares to accessibility by car, which is a much more studied field. Eurostat provides a classification of Local Administrative Units (LAUs) based on whether residents can reach a city of at least 50,000 inhabitants by car within 45 minutes. This classification, which distinguishes between remote and non-remote areas, is publicly available on [Eursotats](https://ec.europa.eu/eurostat/web/nuts/local-administrative-units) website and allows for a consistent comparison across Europe. With this data, the left image of Figure [5.10](#page-102-1) is created. Additionally, the Spanish National Geographic Institute (IGN ) published a study in 2022 (Instituto Geogr´afico Nacional (IGN), [2022\)](#page-112-8) estimating travel times by car from each municipality to the nearest cities of 5,000, 20,000, and 50,000 inhabitants. Making use of the final data set, it was possible to compute the central image of Figure [5.10.](#page-102-1) As can be seen, even if both classifications analyse the same, the different methodologies used generate discrepancies in the results. Finally, and in order to compare these results with those obtained with the methodology described in this thesis, another calculation was made, computing travel times to the two municipalities having more than 50,000 inhabitants in Tarragona: Tarragona city and Reus. This comparison is however not perfect, since the public transport analysis is limited to municipalities in the province of Tarragona, while the IGN and Eurostat data encompass the entire national territory. By comparing these car-based indicators with the results from the proposed public transport methodology, it is noticeable that, as expected, the public transport accessibility and coverage is significantly smaller. This could however be also partly due to the island effect introduced in the bus analysis A second comparison, this time only using IGN data is also presented. Figures [5.11](#page-103-0) to [5.13](#page-104-1) illustrate the comparison between travel times for reaching cities of 5,000, 20,000 and 50,000 inhabitants by car (left) and bus (right). The results confirm a clear advantage in accessibility by car, with travel times being shorter and coverage more comprehensive. The data supports these findings. For example, when comparing travel times by bus and car to municipalities with 5,000 inhabitants or more, data reveals that 34 municipalities have shorter bus travel times compared to car travel times, while 104 municipalities show the opposite trend. However, on average, when bus travel is quicker, the difference is about 3.2 minutes, indicating that travel times are almost identical. Conversely, when bus travel takes longer, the average difference is of 33.5 minutes, meaning that bus travel is significantly slower. A similar analysis was conducted for accessibility to municipalities with 20,000 and 50,000 inhabitants, and the results are presented in Table [5.1.](#page-102-2) ![](_page_102_Figure_1.jpeg) Figure 5.10: Comparison of accessibility by car and public transport to cities with more than 50,000 inhabitants in 45 minutes according to Eursotats (car, left), IGN (car, center) and the methodology of this thesis (bus, right). Table 5.1: Comparison of travel times by bus and car to the closest municipalities with 5,000, 20,000 and 50,000 inhabitants municipalities. These comparisons may provide a hint of why Spain remains a predominantly car centric society, particularly in rural areas. If travel times by car are always faster or as long as those with public transport, people will systematically choose the most comfortable option, and therefore opt for car. # 5.4.2 Assessing municipalities connectivity to their regional heads Another valuable insight that could be drawn from this analysis is whether a municipality's fastest connection to a regional head, is actually to the regional head it officially belongs to. This type of analysis can be useful to identify mismatches in functional connectivity and could serve a role similar to urban-rural catchment area studies, such as those seen in the URCA (Urban-Rural Catchment Areas) framework investigated in Cattaneo et al. [\(2021\)](#page-109-3). This is explored in Figure [5.14.](#page-105-0) In the map, each dark colour represents a regional head, and light colour areas indicate municipalities for which that head is the fastest to reach. For instance, all municipalities in light red have a faster access to the regional head in dark red than to any other regional head. On the other hand the administrative official limits of each region are in bold. Therefore, what would be expectable o desirable is that all municipalities in a same Comarca appear in the same colour, the one of their regional head. As can be seen, in some cases, such as Tortosa, Montblanc, or Valls, this is the case. However, in others, such as Amposta, it can be seen that approximately 40% of its municipalities have faster access to Tortosa than to Amposta itself, their regional head. A similar analysis could be done with other essential facilities such as hospitals or educations institutions, to check if the facility regions are assigned to the essential facilities that ![](_page_103_Figure_2.jpeg) Figure 5.11: Comparison of accessibility to the closest 5,000 inhabitants municipalities by car (left) or bus (right). are more convenient for them. ![](_page_104_Figure_2.jpeg) Figure 5.12: Comparison of accessibility to the closest 20,000 inhabitants municipalities by car (left) or bus (right). ![](_page_104_Figure_5.jpeg) Figure 5.13: Comparison of accessibility to the closest 50,000 inhabitants municipalities by car (left) or bus (right). ![](_page_105_Figure_1.jpeg) Figure 5.14: Comparison between better connected regional head and assigned administrative regional head using bus per municipality. # 5.5 Conclusion Principle 20 of the European Pillar of Social (European Commission, [2017\)](#page-110-4) establishes that everyone has the right to access essential services. To that end, it is required to have an efficient and robust transport network that enables access to these destinations at an affordable price and within a reasonable amount of time. Currently, this has become a priority for the European Commission, which has very recently (22 of May 2025) published a commission recommendation to the member states about this matter (European Commission, [2025\)](#page-111-2). On this official publication, the EU commission invites the member states to develop a strategic approach to prevent and combat transport poverty, including the identification of those in need. It is precisely to help identifying those in need that the current accessibility indicator is developed. This indicator builds on previously existing literature, namely Radics et al. [\(2024\)](#page-114-0), and investigates which essential trips (trips done to essential facilities, that enable full participation of individuals in life) can be done within a reasonable amount of time, directly relating to the accessibility definition proposed by the EU commission (Cludius et al., [2024,](#page-109-0) European Parliament and Council, [2022\)](#page-111-7). In addition, the current work also frames the accessibility dimension within the transport poverty concept, and establishes some ordering of the dimensions (see Figure [3.1\)](#page-45-0). Based on it, accessibility comes as the last dimension to be evaluated, only for those transports deemed available, adequate, and affordable. To measure accessibility in this context, the following data is required. In the first place, it is needed to establish the granularity of the geometric scope to be adopted (province, municipality, neighbourhood, etc.), and to classify the population in several groups (youngsters, elderly, adults, etc.). The second step is to establish, for each of these groups, which are the essential services they need to have access to (namely which are the facilities providing access to them), how frequently (how many times a month) should they have access to them, and what is the maximum time that reaching these locations should take. Then, GTFS or other transportation network data is required to compute the actual time it takes citizens to reach essential services, and evaluate if it is in the below the acceptable threshold. By doing so, it is possible to determine the percentage of essential trips that can be done in a reasonable amount of time, this is accessibility in the context of transport poverty. Finally, the number of citizens in each group and area of study should be included, to provide a wider view of the number of citizens suffering from it. Contextualising this measurement again in the general context of transport poverty, the fact of only including the essential trips in the accessibility analysis, entangles that the only acceptable threshold for accessibility should be 100%, and if it is less, then the individual could be considered to be suffering from transport poverty. However, the percentage measurement also allows to nuance and acknowledge that not all citizens may suffer from poverty with the same severity. To demonstrate its applicability in real life scenarios, an example case study in Tarragona is presented. Observing the results, it can be seen that the current indicator puts in the spotlight accessibility, and places significant efforts in presenting the information in a way that it becomes relevant for decision makers. Unlike other recent accessibility-related indicators, such as those presented in Instituto Geogr´afico Nacional (IGN) [\(2022\)](#page-112-8) or Cludius et al. [\(2024\)](#page-109-0), the proposed methodology offers accessibility as a unique measure that comprehends all the calculations associated with it, and outputs directly usable information, pinpointing precisely what are the services each municipality lacks access to. If the results are directly taken as reliable, this work outputs that 85.6% of the rural municipalities in Tarragona (107 out of 125) suffer from transport poverty. However, this methodology provides precise information about the causes provoking this situation. For instance, the results would recommend to re-design the network in Tarragona to provide better access to hospitals, particularly to the 64 municipalities that do not have good enough connexion to them. Similarly it would be recommended to invest in leisure areas in the northern half of the region, as all of them are concentrated near the coast, increasing travel times for northern regions. Finally, the biggest effort should be put in ensuring that all municipalities can access their regional heads in less than half an hour. This would greatly increase accessibility for adults, that currently have lower accessibility levels than elderly and youngsters. The proposed methodology is generic, but implies a number of choices. In the example presented, these choices are made aiming to output results that are meaningful and aligned with the European directives and the Spanish aims. This is, to prove useful for decision makers. Firstly, this work makes the choice of studying accessibility by public transport, a less explored area when compared to accessibility by car (see Instituto Geogr´afico Nacional (IGN), [2022,](#page-112-8) Eurostat, [2025\)](#page-111-0), which is crucial for sustainability. This choice deviates attention from cars, and places it in promoting public transport, which is a less polluting way of transport (Ritchie, [2023\)](#page-114-3). This aligns directly with the ONU Sustainable Development Goals, the European Green Deal aim and the European and Spanish mobility strategy (see Section [2.3.2\)](#page-39-0). Additionally, the choices made in the result section concerning the geometric granularity levels, make them more reliable for rural areas, which are considered vulnerable in terms of transport poverty (European Parliament and Council, [2022\)](#page-111-7). This is mainly due to the computation method, that allows to compute times between municipalities (good for rural areas) but not within them (suited for urban areas), and to the inclusion of only intermunicipal GTFS data, not including urban networks such as tram, metro, urban buses or even active mobility. The aim is thus to ensure inclusivity, identifying those most in need, for the ultimate end that no one is left behind (European Commission, [2021c\)](#page-111-5). Further aligning with EU ambitions, all data used is open source, encouraging its use and demonstrating the valuable indicators that can be derived from integrating different data sources. Furthermore, the adopted thresholds align with Spain's ambition to become a "30-minute country," ensuring all services are accessible by public transport. Apart from that, this vision also aligns with a transport equity view, in which the goal is to "replace the traditional measure of travel time savings that favour better-off societal groups who are travelling more, with accessibility gains measures that cater for more vulnerable social groups" (Di Ciommo and Shiftan, [2017,](#page-110-15) p.140). At the end of the document, the discussion section also enriches the debate, opening up the door for new venues and ways to refine the proposed methodology. In particular, several proposals on how to account for the number of citizens affected, or how to modify the compliance parameters are proposed. Deviating slightly from the main purpose of the work, this study also briefly presents that in general, accessibility by public transport is lower than accessibility by car (see Section [5.4.1.](#page-101-1) This could partly explain why Spain remains largely a car centric country, where still in 2024, according to the [Microdata of 2024 from](https://ine.es/dyngs/INEbase/operacion.htm?c=Estadistica_C&cid=1254736176990&menu=resultados&idp=1254735576863#_tabs-1254736195369) [INE, Encuesta de Turismo de Residentes ETR](https://ine.es/dyngs/INEbase/operacion.htm?c=Estadistica_C&cid=1254736176990&menu=resultados&idp=1254735576863#_tabs-1254736195369) 78,9% of the trips where made in a private owned car. This view aligns with available statistics results, such as EMEF study that in the page 60 of the results of 2023, show that 39 % of people that do not take public transport because the offer is inadequate for their purposes (Metropoli, [2023\)](#page-113-13). To sum up, this work shows that the accessibility dimension of transport poverty can be quantified and summarized in a way that it is useful for decision makers, and that should help them when allocating transport investments. In addition it offers guidance on how to evaluate and include accessibility in the context of transport poverty, proposing a new lens to evaluate the relation between the different dimensions. This documents also proposes an indicator for measuring accessibility, based on previously existing literature, but highlighting its focus transport poverty, by only including essential trips to this evaluation. Finally, a study case is offered, demonstrating that this methodology can prove useful when evaluating transport poverty in rural areas. It therefore answers all research questions. # Bibliography - Alguacil Denche, A., Romero Mora, J. C., P´erez Bravo, M., & Collado Van-Baumberghen, N. (2024). Vulnerabilidad y pobreza en el transporte en espa˜na. ECODES. - Allen, J., & Farber, S. (2019). Sizing up transport poverty: A national scale accounting of low-income households suffering from inaccessibility in canada, and what to do about it. Transport Policy, 74, 214–223. [https://doi.org/https://doi.org/10.1016/j.tranpol.](https://doi.org/https://doi.org/10.1016/j.tranpol.2018.11.018) [2018.11.018](https://doi.org/https://doi.org/10.1016/j.tranpol.2018.11.018) - Arasan, V. T., Wermuth, M., & Srinivas, B. (1996). Modeling of stratified urban trip distribution. Journal of Transportation Engineering-asce, 122, 342–349. [https://doi.org/](https://doi.org/10.1061/(ASCE)0733-947X(1996)122:5(342)) [10.1061/\(ASCE\)0733-947X\(1996\)122:5\(342\)](https://doi.org/10.1061/(ASCE)0733-947X(1996)122:5(342)) - Association, C. U. T., et al. (2004). Bus rapid transit: A canadian perspective. Issues Paper, 10. - Beir˜ao, G., & Sarsfield Cabral, J. (2007). Understanding attitudes towards public transport and private car: A qualitative study. Transport Policy, 14 (6), 478–489. [https://doi.](https://doi.org/https://doi.org/10.1016/j.tranpol.2007.04.009) [org/https://doi.org/10.1016/j.tranpol.2007.04.009](https://doi.org/https://doi.org/10.1016/j.tranpol.2007.04.009) - Bosch, R. M. (2025, April 1). Radiograf´ıa de la espa˜na vac´ıa para avanzar hacia el "pa´ıs de los 30 minutos" [Accessed: 2025-04-24]. - Buliung, R. N., & Kanaroglou, P. S. (2007). Activity–travel behaviour research: Conceptual issues, state of the art, and emerging perspectives on behavioural analysis and simulation modelling. Transport Reviews, 27 (2), 151–187. [https: / / doi. org / 10. 1080 /](https://doi.org/10.1080/01441640600858649) [01441640600858649](https://doi.org/10.1080/01441640600858649) - C´amara, A., Mart´ınez, M. I., & Santero-S´anchez, R. (2020). Macroeconomic cost of excluding persons with disabilities from the workforce in spain. IZA Journal of Labor Policy, 10 (1). - Capasso Da Silva, D., King, D. A., & Lemar, S. V. (2019). Accessibility in practice: 20minute city as a sustainability planning goal. Sustainability, 12 (1), 129. [https://doi.org/10.](https://doi.org/10.3390/su12010129) [3390/su12010129](https://doi.org/10.3390/su12010129) - Carlos, P., & Director, M. (2020). Urban and territorial transitions [Accessed: 2025-04-24]. - Carruthers, R., Dick, M., & Saurkar, A. (2005). Affordability of public transport in developing countries (tech. rep.). The World Bank Group. - Cattaneo, A., Nelson, A., & McMenomy, T. (2021). Global mapping of urban–rural catchment areas reveals unequal access to services. Proceedings of the National Academy of Sciences, 118 (2), e2011990118. - Chu, Y.-L. (1990). Combined trip distribution and assignment model incorporating captive travel behavior. TRANSPORTATION RESEARCH RECORD 1285. - Cludius, J., Noka, V., Unger, N., Delfosse, L., Dolinga, T., Schumacher, K., Suta, C.-M., Lechtenfeld, R., Vornicu, A., Sinea, A., et al. (2024). Transport Poverty: Definitions, Indicators, Determinants, and Mitigation Strategies (tech. rep.). European Comission. - [https : / / employment social affairs . ec . europa . eu / transport poverty definitions](https://employment-social-affairs.ec.europa.eu/transport-poverty-definitions-indicators-determinants-and-mitigation-strategies-final-report_en) [indicators-determinants-and-mitigation-strategies-final-report](https://employment-social-affairs.ec.europa.eu/transport-poverty-definitions-indicators-determinants-and-mitigation-strategies-final-report_en) en - Codina, E. (2025). Entropy-based trip distribution models [Accessed: 2025-05-26]. [https://](https://upcommons.upc.edu/bitstream/handle/2117/425891/distr-models-entropia.pdf?sequence=1) [upcommons.upc.edu/bitstream/handle/2117/425891/distr - models - entropia.pdf ?](https://upcommons.upc.edu/bitstream/handle/2117/425891/distr-models-entropia.pdf?sequence=1) [sequence=1](https://upcommons.upc.edu/bitstream/handle/2117/425891/distr-models-entropia.pdf?sequence=1) - Correa, J., & Stier-Moses, N. (2011, February). Wardrop equilibria. [https://doi.org/10.1002/](https://doi.org/10.1002/9780470400531.eorms0962) [9780470400531.eorms0962](https://doi.org/10.1002/9780470400531.eorms0962) - Ley de Ordenaci´on de los Transportes Terrestres (1987). [https://www.boe.es/buscar/act.](https://www.boe.es/buscar/act.php?id=BOE-A-1987-17803) [php?id=BOE-A-1987-17803](https://www.boe.es/buscar/act.php?id=BOE-A-1987-17803) - Ley 45/2007, de 13 de diciembre, para el desarrollo sostenible del medio rural (2007, December). - The Revised European Urban Charter - CPL (11) 7 Part II (2008). [https: / / rm.coe.int /](https://rm.coe.int/090000168071a7e9) [090000168071a7e9](https://rm.coe.int/090000168071a7e9) - Joint Report by the Commission and the Council on Social Inclusion (2003). [https: / /ec.](https://ec.europa.eu/employment_social/soc-prot/soc-incl/final_joint_inclusion_report_2003_en.pdf) europa.eu/employment [social/soc-prot/soc-incl/final](https://ec.europa.eu/employment_social/soc-prot/soc-incl/final_joint_inclusion_report_2003_en.pdf) joint inclusion report 2003 en. [pdf](https://ec.europa.eu/employment_social/soc-prot/soc-incl/final_joint_inclusion_report_2003_en.pdf) - Crisp, R., Gore, T., & McCarthy, L. (2017). Addressing transport barriers to work in low income neighbourhoods: A review of evidence and practice (Project Report). Sheffield Hallam University. Sheffield. - Daganzo, C. (2014). Multinomial probit: The theory and its application to demand forecasting. Elsevier. - de Grange, L., Troncoso, R., & Gonz´alez, F. (2012). An empirical evaluation of the impact of three urban transportation policies on transit use. Transport Policy, 22, 11–19. - Delmelle, E. C., & Casas, I. (2012). Evaluating the spatial equity of bus rapid transit-based accessibility patterns in a developing country: The case of cali, colombia [URBAN TRANSPORT INITIATIVES]. Transport Policy, 20, 36–46. [https://doi.org/https:](https://doi.org/https://doi.org/10.1016/j.tranpol.2011.12.001) [//doi.org/10.1016/j.tranpol.2011.12.001](https://doi.org/https://doi.org/10.1016/j.tranpol.2011.12.001) - Di Ciommo, F., & Shiftan, Y. (2017). Transport equity analysis. Transport Reviews, 37 (2), 139–151. - Enderami, S., Sutley, E., Helgeson, J., et al. (2024). Measuring post-disaster accessibility to essential goods and services: Proximity, availability, adequacy, and acceptability dimensions. Journal of Infrastructure Preservation and Resilience, 5, 12. [https://doi.](https://doi.org/10.1186/s43065-024-00104-0) [org/10.1186/s43065-024-00104-0](https://doi.org/10.1186/s43065-024-00104-0) - Etminani-Ghasrodashti, R., & Ardeshiri, M. (2015). Modeling travel behavior by the structural relationships between lifestyle, built environment and non-working trips. Transportation Research Part A-policy and Practice, 78, 506–518. [https://doi.org/10.1016/](https://doi.org/10.1016/J.TRA.2015.06.016) [J.TRA.2015.06.016](https://doi.org/10.1016/J.TRA.2015.06.016) - The European pillar of social rights in 20 principles (2017). [https: / / employment - social](https://employment-social-affairs.ec.europa.eu/european-pillar-social-rights-20-principles_en) [affairs.ec.europa.eu/european-pillar-social-rights-20-principles](https://employment-social-affairs.ec.europa.eu/european-pillar-social-rights-20-principles_en) en - European Commission. (2021a). Eu emissions trading system (eu ets) and the social climate fund [Video]. [https://audiovisual.ec.europa.eu/corporateplayer/index.html?video=I-](https://audiovisual.ec.europa.eu/corporateplayer/index.html?video=I-256989&language=EN)[256989&language=EN](https://audiovisual.ec.europa.eu/corporateplayer/index.html?video=I-256989&language=EN) - Regulation (EU) 2021/1119 of the European Parliament and of the Council of 30 June 2021 establishing the framework for achieving climate neutrality and amending Regulations (EC) No 401/2009 and (EU) 2018/1999 (2021). [http://data.europa.eu/eli/reg/2021/](http://data.europa.eu/eli/reg/2021/1119/oj) [1119/oj](http://data.europa.eu/eli/reg/2021/1119/oj) European Commission. (2021c). Sustainable and smart mobility strategy ((Accessed: 07 February 2025)). Mobility and Transport. [https: / / transport.ec.europa.eu / transport](https://transport.ec.europa.eu/transport-themes/mobility-strategy_en) [themes/mobility-strategy](https://transport.ec.europa.eu/transport-themes/mobility-strategy_en) en - Commission Recommendation of 22.5.2025 on Transport Poverty: Ensuring Affordable, Accessible and Fair Mobility (2025). [https://eur-lex.europa.eu/legal-content/EN/TXT/](https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=OJ:L_202501021) [PDF/?uri=OJ:L](https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=OJ:L_202501021) 202501021 - European Commission and Directorate-General for Mobility and Transport. (2011). White paper on transport – roadmap to a single european transport area – towards a competitive and resource efficient transport system. Publications Office of the European Union. - Fit for 55: Delivering the EU's 2030 Climate Target on the Way to Climate Neutrality (2021). - European Environment Agency. (2025, February). Transport and mobility [Accessed: 2025- 02-20]. - Directive 2003/87/EC of the European Parliament and of the Council of 13 October 2003 establishing a scheme for greenhouse gas emission allowance trading within the Community and amending Council Directive 96/61/EC (2003). [https://eur-lex.europa.](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:02003L0087-20230605) [eu/legal-content/EN/TXT/?uri=CELEX:02003L0087-20230605](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:02003L0087-20230605) - The European Green Deal (2019). [https: / / commission. europa. eu / strategy - and - policy /](https://commission.europa.eu/strategy-and-policy/priorities-2019-2024/european-green-deal_en) [priorities-2019-2024/european-green-deal](https://commission.europa.eu/strategy-and-policy/priorities-2019-2024/european-green-deal_en) en - Social Climate Fund (2022, June). [https://www.europarl.europa.eu/doceo/document/TA-9-](https://www.europarl.europa.eu/doceo/document/TA-9-2022-0247_EN.html) [2022-0247](https://www.europarl.europa.eu/doceo/document/TA-9-2022-0247_EN.html) EN.html - Regulation 2023/955 - en - EUR-lex (2024). [https://eur-lex.europa.eu/legal-content/EN/](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32023R0955) [TXT/?uri=CELEX%3A32023R0955](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32023R0955) - Eurostat. (2021). Territorial typologies manual - cluster types [Statistics Explained, Eurostat]. [https: / / ec. europa. eu / eurostat / statistics - explained /index. php ? title=Territorial](https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Territorial_typologies_manual_-_cluster_types) [typologies](https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Territorial_typologies_manual_-_cluster_types) manual - cluster types - Eurostat. (2025). Statistics explained [Accessed: 2025-05-13]. [https://ec.europa.eu/eurostat/](https://ec.europa.eu/eurostat/statistics-explained/index.php?oldid=657242) [statistics-explained/index.php?oldid=657242](https://ec.europa.eu/eurostat/statistics-explained/index.php?oldid=657242) - Farinloye, T., Mogaji, E., Aririguzoh, S., & Kieu, T. A. (2019). Qualitatively exploring the effect of change in the residential environment on travel behaviour. Travel Behaviour and Society, 17, 26–35. - Frederick, C., & Gilderbloom, J. (2018). Commute mode diversity and income inequality: An inter-urban analysis of 148 midsize us cities. Local Environment, 23 (1), 54–76. - Furness, K. P. (1965). Time function iteration. Traffic Engineering and Control, 7 (7), 458– 460. - Gil Sol´a, A., & Vilhelmson, B. (2019). Negotiating proximity in sustainable urban planning: A swedish case. Sustainability, 11 (1). - Real Decreto 1211/1990, de 28 de septiembre, por el que se aprueba el Reglamento de la Ley de Ordenaci´on de los Transportes Terrestres (1990). [https://www.boe.es/buscar/act.](https://www.boe.es/buscar/act.php?id=BOE-A-1990-24442) [php?id=BOE-A-1990-24442](https://www.boe.es/buscar/act.php?id=BOE-A-1990-24442) - Espa˜na 2050. Fundamentos y propuestas para una Estrategia Nacional de Largo Plazo. Ministerio de la Presidencia (2021). [https://www.lamoncloa.gob.es/presidente/actividades/](https://www.lamoncloa.gob.es/presidente/actividades/Documents/2021/200521-Estrategia_Espana_2050.pdf) [Documents/2021/200521-Estrategia](https://www.lamoncloa.gob.es/presidente/actividades/Documents/2021/200521-Estrategia_Espana_2050.pdf) Espana 2050.pdf Plan de Recuperaci´on: 130 medidas frente al reto demogr´afico (2021). [https://www.%20miteco.](https://www.%20miteco.%20gob.%20es/es/reto-demografico/temas/plan%5C_recuperacion%5C_130%5C_medidas%5C_tcm30-524369.pdf) [%20gob . %20es / es / reto - demografico / temas / plan % 5C](https://www.%20miteco.%20gob.%20es/es/reto-demografico/temas/plan%5C_recuperacion%5C_130%5C_medidas%5C_tcm30-524369.pdf) recuperacion % 5C 130 % 5C medidas%5C [tcm30-524369.pdf](https://www.%20miteco.%20gob.%20es/es/reto-demografico/temas/plan%5C_recuperacion%5C_130%5C_medidas%5C_tcm30-524369.pdf) - Gobierno de Espa˜na. (2021c, December). La poblaci´on de las ´areas rurales en espa˜na supera los 7,5 millones de personas [Accessed: 2025-02-20]. [https://www.mapa.gob.es/es/](https://www.mapa.gob.es/es/prensa/ultimas-noticias/detalle_noticias.aspx?tcm=tcm:30-583990) prensa/ultimas-noticias/detalle [noticias.aspx?tcm=tcm:30-583990](https://www.mapa.gob.es/es/prensa/ultimas-noticias/detalle_noticias.aspx?tcm=tcm:30-583990) - Gobierno de Espa˜na. (2024). La tasa bruta de graduaci´on en bachillerato se sit´ua en el 75,4% en el curso 2022-2023 [Acceso el 6 de mayo de 2025, Ministerio de Educaci´on, Formaci´on Profesional y Deportes]. [https://www.educacionfpydeportes.gob.es/prensa/](https://www.educacionfpydeportes.gob.es/prensa/actualidad/2024/12/20241220-anuarioestadistico.html) [actualidad/2024/12/20241220-anuarioestadistico.html](https://www.educacionfpydeportes.gob.es/prensa/actualidad/2024/12/20241220-anuarioestadistico.html) - GTFS.org. (n.d.). General transit feed specification (gtfs) [Accessed: 2025-05-08]. - Guagliardo, M. F. (2004). Spatial accessibility of primary care: Concepts, methods and challenges. International Journal of Health Geographics, 3 (1), 3. [https://doi.org/10.1186/](https://doi.org/10.1186/1476-072X-3-3) [1476-072X-3-3](https://doi.org/10.1186/1476-072X-3-3) - Handy, S., Cao, X., & Mokhtarian, P. (2005). Correlation or causality between the built environment and travel behavior? evidence from northern california. Transportation Research Part D: Transport and Environment, 10 (6), 427–444. [https://doi.org/https:](https://doi.org/https://doi.org/10.1016/j.trd.2005.05.002) [//doi.org/10.1016/j.trd.2005.05.002](https://doi.org/https://doi.org/10.1016/j.trd.2005.05.002) - Hansen, W. G. (1959). How accessibility shapes land use. Journal of the American Institute of Planners, 25 (2), 73–76. - Hanson, S., & Hanson, P. O. (1981). The travel-activity patterns of urban residents: Dimensions and relationships to sociodemographic characteristics. Economic Geography, 57, 332–347. - Instituto Geogr´afico Nacional (IGN). (2022). C´alculo de distancias a trav´es de la red viaria de la igr-rt [Accessed: 2025-05-25]. [https : / / storymaps . arcgis . com / stories /](https://storymaps.arcgis.com/stories/662be0691ccd4d6f8d134817c409a356) [662be0691ccd4d6f8d134817c409a356](https://storymaps.arcgis.com/stories/662be0691ccd4d6f8d134817c409a356) - Instituto Nacional de Estad´ıstica. (2025). Contabilidad regional de espa˜na. resultados 2023 [Accessed: 2023-05-08]. [https : / / ine . es / dyngs / INEbase / es / operacion . htm ? c =](https://ine.es/dyngs/INEbase/es/operacion.htm?c=Estadistica_C&cid=1254736167628&menu=resultados&idp=1254735576581) Estadistica [C&cid=1254736167628&menu=resultados&idp=1254735576581](https://ine.es/dyngs/INEbase/es/operacion.htm?c=Estadistica_C&cid=1254736167628&menu=resultados&idp=1254735576581) - Instituto Nacional de Estad´ıstica (INE). (2024). Poblaci´on de 16 y m´as a˜nos seg´un nivel de formaci´on alcanzado, sexo y grupo de edad [Acceso el 6 de mayo de 2025]. [https:](https://www.ine.es/jaxiT3/Datos.htm?t=6369) [//www.ine.es/jaxiT3/Datos.htm?t=6369](https://www.ine.es/jaxiT3/Datos.htm?t=6369) - Jaramillo, C., Liz´arraga, C., & Grindlay, A. L. (2012). Spatial disparity in transport social needs and public transport provision in santiago de cali (colombia) [Special Section on Theoretical Perspectives on Climate Change Mitigation in Transport]. Journal of Transport Geography, 24, 340–357. [https: / / doi. org / https: / / doi. org / 10. 1016 / j.](https://doi.org/https://doi.org/10.1016/j.jtrangeo.2012.04.014) [jtrangeo.2012.04.014](https://doi.org/https://doi.org/10.1016/j.jtrangeo.2012.04.014) - Kampert, A., Nijenhuis, J., Nijland, H., Uitbeijerse, G., & Verhoeven, M. (2019). Indicator risico op vervoersarmoede (tech. rep.). CBS. [https: / /www.%20cbs.%20nl / nl - nl /](https://www.%20cbs.%20nl/nl-nl/achtergrond/2019/42/indicator-risico-op-vervoersarmoede) [achtergrond/2019/42/indicator-risico-op-vervoersarmoede](https://www.%20cbs.%20nl/nl-nl/achtergrond/2019/42/indicator-risico-op-vervoersarmoede) - Kampert, A., Verhoeven, M., Nijenhuis, J., & Dahlmans, D. (2018). Risico op vervoersarmoede een eerste aanzet tot een indicator (tech. rep.). CBS. Retrieved from. - Kenyon, S., Lyons, G., & Rafferty, J. (2002). Transport and social exclusion: Investigating the possibility of promoting inclusion through virtual mobility. Journal of Transport Geography, 10 (3), 207–219. [https://doi.org/https://doi.org/10.1016/S0966-6923\(02\)](https://doi.org/https://doi.org/10.1016/S0966-6923(02)00012-1) [00012-1](https://doi.org/https://doi.org/10.1016/S0966-6923(02)00012-1) Kiss, M. (2022, October). Understanding transport poverty (PE 738.181) (Accessed: 2025-02- 20). European Parliamentary Research Service. [https://www.europarl.europa.eu/](https://www.europarl.europa.eu/RegData/etudes/ATAG/2022/738181/EPRS_ATA(2022)738181_EN.pdf) [RegData/etudes/ATAG/2022/738181/EPRS](https://www.europarl.europa.eu/RegData/etudes/ATAG/2022/738181/EPRS_ATA(2022)738181_EN.pdf) ATA(2022)738181 EN.pdf - Kompil, M., Jacobs-Crisioni, C., Perpi˜n´a Castillo, C., & Lavalle, C. (2022). Accessibility to services in europe's member states – an evaluation by degree of urbanisation and remoteness. - Kong, W., Pojani, D., Sipe, N., & Stead, D. (2021). Transport poverty in chinese cities: A systematic literature review. Sustainability, 13 (9). - Litman, T. (2017). Evaluating transportation equity. Victoria Transport Policy Institute Victoria, BC, Canada. - Llorens, A. (2020). La desaparici´on del comercio en los pueblos de tarragona debido a la despoblaci´on [Consultado el 21 de mayo de 2025]. La Vanguardia. [https : / / www .](https://www.lavanguardia.com/local/tarragona/20200124/472888548283/comercio-tiendas-pueblos-despoblacion-pimec-tarragona.html) [lavanguardia . com / local / tarragona / 20200124 / 472888548283 / comercio - tiendas](https://www.lavanguardia.com/local/tarragona/20200124/472888548283/comercio-tiendas-pueblos-despoblacion-pimec-tarragona.html) [pueblos-despoblacion-pimec-tarragona.html](https://www.lavanguardia.com/local/tarragona/20200124/472888548283/comercio-tiendas-pueblos-despoblacion-pimec-tarragona.html) - Logan, T., & Guikema, S. (2020). Reframing resilience: Equitable access to essential services. Risk Analysis, 40. - Lucas, K. (2012). Transport and social exclusion: Where are we now? [URBAN TRANSPORT INITIATIVES]. Transport Policy, 20, 105–113. [https://doi.org/https://doi.org/10.](https://doi.org/https://doi.org/10.1016/j.tranpol.2012.01.013) [1016/j.tranpol.2012.01.013](https://doi.org/https://doi.org/10.1016/j.tranpol.2012.01.013) - Lucas, K. (2018). Editorial for special issue of european transport research review: Transport poverty and inequalities. European Transport Research Review, 10, 17. [https://doi.](https://doi.org/10.1007/s12544-018-0288-6) [org/10.1007/s12544-018-0288-6](https://doi.org/10.1007/s12544-018-0288-6) - Mann, E., & Abraham, C. (2006). The role of affect in uk commuters' travel mode choices: An interpretative phenomenological analysis. British journal of psychology, 97 (2), 155– 176. - Maucorps, A., Mart´ın, J. C., Dentinho, T. L. C. P., Verg´e-D´epr´e, C. R., Peral, P. P., & Fortuna, M. J. A. (2025). Transport and tourism in outermost regions: Assessing mobility poverty and the effects of new climate policies (tech. rep.). European Parliament, Policy Department for Transport, Employment and Social Affairs. [https :](https://www.europarl.europa.eu/RegData/etudes/STUD/2025/759311/CASP_STU(2025)759311_EN.pdf) [//www.europarl.europa.eu/RegData/etudes/STUD/2025/759311/CASP](https://www.europarl.europa.eu/RegData/etudes/STUD/2025/759311/CASP_STU(2025)759311_EN.pdf) STU(2025) 759311 [EN.pdf](https://www.europarl.europa.eu/RegData/etudes/STUD/2025/759311/CASP_STU(2025)759311_EN.pdf) - McPhearson, T., Haase, D., Kabisch, N., & Gren, ˚A. (2016). Advancing understanding of the complex nature of urban systems [Navigating Urban Complexity: Advancing Understanding of Urban Social – Ecological Systems for Transformation and Resilience]. Ecological Indicators, 70, 566–573. [https://doi.org/https://doi.org/10.1016/j.ecolind.](https://doi.org/https://doi.org/10.1016/j.ecolind.2016.03.054) [2016.03.054](https://doi.org/https://doi.org/10.1016/j.ecolind.2016.03.054) - McQuaid, R. W. (2009). A model of the travel to work limits of parents [Symposium on Transport and Particular Populations]. Research in Transportation Economics, 25 (1), 19–28. - Metropoli, I. (2023, April). Enquesta de mobilitat en dia feiner 2023 (emef 2023) - resum executiu [Accessed: 2025-06-01]. [https://ce- sermetra.atm.cat/documents/662112/](https://ce-sermetra.atm.cat/documents/662112/1628687/EMEF2023_ResumExecutiu.pdf/52251169-60c4-7bfd-16e6-049766b46177?t=1729230529198) 1628687/EMEF2023 [ResumExecutiu.pdf/52251169- 60c4- 7bfd- 16e6- 049766b46177?](https://ce-sermetra.atm.cat/documents/662112/1628687/EMEF2023_ResumExecutiu.pdf/52251169-60c4-7bfd-16e6-049766b46177?t=1729230529198) [t=1729230529198](https://ce-sermetra.atm.cat/documents/662112/1628687/EMEF2023_ResumExecutiu.pdf/52251169-60c4-7bfd-16e6-049766b46177?t=1729230529198) - Metta, P. (2020). Transport poverty in thailand: Concept, measurement and data availability. International Review for Spatial Planning and Sustainable Development, 8 (2), 70–85. [https://doi.org/10.14246/irspsd.8.2](https://doi.org/10.14246/irspsd.8.2_70) 70 Meyer, M. D., & Miller, E. J. (2001). Urban transportation planning: A decision-oriented approach. - Ministerio de Sanidad. (2021). Tablas ccaa 2021 [Accedido: 2025-05-09, Ministerio de Sanidad]. [https: / /www. sanidad. gob. es / estadEstudios / estadisticas / docs /TablasSIAE2021 /](https://www.sanidad.gob.es/estadEstudios/estadisticas/docs/TablasSIAE2021/Tablas_CCAA_2021.pdf) Tablas CCAA [2021.pdf](https://www.sanidad.gob.es/estadEstudios/estadisticas/docs/TablasSIAE2021/Tablas_CCAA_2021.pdf) - Ministerio de Transportes y Movilidad Sostenible. (2021). Estrategia de movilidad segura, sostenible y conectada 2030 ((Accessed: 07 February 2025)). Ministerio de Transportes y Movilidad Sostenible, Espa˜na. [https://www.transportes.gob.es/ministerio/planes](https://www.transportes.gob.es/ministerio/planes-estrategicos/esmovilidad)[estrategicos/esmovilidad](https://www.transportes.gob.es/ministerio/planes-estrategicos/esmovilidad) - Morency, C., Paez, A., Roorda, M. J., Mercado, R., & Farber, S. (2011). Distance traveled in three canadian cities: Spatial analysis from the perspective of vulnerable population segments. Journal of Transport Geography, 19 (1), 39–50. [https : / / doi . org / https :](https://doi.org/https://doi.org/10.1016/j.jtrangeo.2009.09.013) [//doi.org/10.1016/j.jtrangeo.2009.09.013](https://doi.org/https://doi.org/10.1016/j.jtrangeo.2009.09.013) - National Cooperative Highway Research Program. (n.d.). Chapter 5: Random utility models [Accessed: 2025-05-31]. [https://onlinepubs.trb.org/onlinepubs/nchrp/cd-22/manual/](https://onlinepubs.trb.org/onlinepubs/nchrp/cd-22/manual/v2chapter5.pdf) [v2chapter5.pdf](https://onlinepubs.trb.org/onlinepubs/nchrp/cd-22/manual/v2chapter5.pdf) - OTLE. (2025, April). Monogr´afico sobre la pobreza de transporte (tech. rep.) (Available at: [https : / / cdn . transportes . gob . es / portal - web - drupal / OTLE / elementos](https://cdn.transportes.gob.es/portal-web-drupal/OTLE/elementos_otle/monografico-pobreza-de-transporte-(abril-2025).pdf) otle / [monografico-pobreza-de-transporte-\(abril-2025\).pdf\)](https://cdn.transportes.gob.es/portal-web-drupal/OTLE/elementos_otle/monografico-pobreza-de-transporte-(abril-2025).pdf). Observatorio del Transporte y la Log´ıstica en Espa˜na (OTLE). - Pappalardo, L., & Simini, F. (2018). Data-driven generation of spatio-temporal routines in human mobility. Data Mining and Knowledge Discovery, 32, 787–829. [https://doi.](https://doi.org/10.1007/s10618-017-0548-4) [org/10.1007/s10618-017-0548-4](https://doi.org/10.1007/s10618-017-0548-4) - para la Transici´on Ecol´ogica y el Reto Demogr´afico, M. (2025). Nuevo marco estrat´egico para la equidad territorial y el reto demogr´afico. [https://www.miteco.gob.es/es/reto](https://www.miteco.gob.es/es/reto-demografico/temas/nuevo-marco-estrategico.html#1_-estrategia-nacional-para-la-equidad-territorial-y-el-reto-demogr%C3%A1fico) [demografico/temas/nuevo-marco-estrategico.html#1](https://www.miteco.gob.es/es/reto-demografico/temas/nuevo-marco-estrategico.html#1_-estrategia-nacional-para-la-equidad-territorial-y-el-reto-demogr%C3%A1fico) -estrategia-nacional-para-la[equidad-territorial-y-el-reto-demogr%C3%A1fico](https://www.miteco.gob.es/es/reto-demografico/temas/nuevo-marco-estrategico.html#1_-estrategia-nacional-para-la-equidad-territorial-y-el-reto-demogr%C3%A1fico) - Pereira, R. H., Karner, A., et al. (2021). Transportation equity. Elsevier. - Pozoukidou, G., & Chatziyiannaki, Z. (2021). 15-minute city: Decomposing the new urban planning eutopia. Sustainability, 13, 928. - Radics, M., Christidis, P., Alonso, B., & dell'Olio, L. (2024). The x-minute city: Analysing accessibility to essential daily destinations by active mobility in seville. Land, 13 (10). - Renfe-Operadora (Grupo). (2023). Presupuestos generales del estado - renfe-operadora (grupo) [[Accessed: 07 February 2025]]. [https: / / www. renfe. com / content / dam / renfe / es /](https://www.renfe.com/content/dam/renfe/es/Grupo-Empresa/Gobierno-corporativo-y-transparencia/Transparencia/pdf/PGE%2023%20Grupo%20Renfe.pdf) [Grupo-Empresa/Gobierno-corporativo-y- transparencia/Transparencia/pdf/PGE%](https://www.renfe.com/content/dam/renfe/es/Grupo-Empresa/Gobierno-corporativo-y-transparencia/Transparencia/pdf/PGE%2023%20Grupo%20Renfe.pdf) [2023%20Grupo%20Renfe.pdf](https://www.renfe.com/content/dam/renfe/es/Grupo-Empresa/Gobierno-corporativo-y-transparencia/Transparencia/pdf/PGE%2023%20Grupo%20Renfe.pdf) - Ritchie, H. (2023). Which form of transport has the smallest carbon footprint? Our World in Data. - Resoluci´on de 28 de febrero de 2019, de la Secretar´ıa de Estado de Funci´on P´ublica, por la que se dictan instrucciones sobre jornada y horarios de trabajo del personal al servicio de la Administraci´on General del Estado y sus organismos p´ublicos (2019). - Sinha, R., Olsson, L. E., & Frostell, B. (2019). Sustainable personal transport modes in a life cycle perspective—public or private? Sustainability, 11 (24). [https://www.mdpi.com/](https://www.mdpi.com/2071-1050/11/24/7092) [2071-1050/11/24/7092](https://www.mdpi.com/2071-1050/11/24/7092) - Stopher, P. R., & McDonald, K. G. (1983). Trip generation by cross-classification: An alternative methodology. Transportation Research Record, 944, 84–91. Sustrans. (2016). Transport poverty in scotland (tech. rep.) (Retrieved from [https://www.](https://www.sustrans.org.uk/media/2880/transport_poverty_in_scotland_2016.pdf) [sustrans.org.uk/media/2880/transport](https://www.sustrans.org.uk/media/2880/transport_poverty_in_scotland_2016.pdf) poverty in scotland 2016.pdf). Sustrans. - Union des Transports Publics et Ferroviaires (UTPF). (2024). Dossier de presse: Observatoire de la mobilit´e attentes du secteur des transports publics urbains et ferroviaires. [https:](https://www.utpf-mobilites.fr/system/files/2024-10/dp_utpf_2024_ok_web.pdf) [//www.utpf-mobilites.fr/system/files/2024-10/dp](https://www.utpf-mobilites.fr/system/files/2024-10/dp_utpf_2024_ok_web.pdf) utpf 2024 ok web.pdf - United Nations Economic Commission for Europe. (2017). Older persons in rural and remote areas. United Nations. - Transforming our world: the 2030 Agenda for Sustainable Development (2015). [https://docs.](https://docs.un.org/en/A/RES/70/1) [un.org/en/A/RES/70/1](https://docs.un.org/en/A/RES/70/1) - Urbanek, A. (2021). Potential of modal shift from private cars to public transport: A survey on the commuters' attitudes and willingness to switch – a case study of silesia province, poland [Policies prompting sustainable transport in urban areas]. Research in Transportation Economics, 85, 101008. [https://doi.org/https://doi.org/10.1016/j.](https://doi.org/https://doi.org/10.1016/j.retrec.2020.101008) [retrec.2020.101008](https://doi.org/https://doi.org/10.1016/j.retrec.2020.101008) - Van Acker, V., Van Wee, B., & Witlox, F. (2010). When transport geography meets social psychology: Toward a conceptual model of travel behaviour. Transport reviews, 30 (2), 219–240. - voor de Leefomgeving, P. (2022, October). Toegang voor iedereen? (An analysis of the accessibility of facilities and jobs in the Netherlands). Planbureau voor de Leefomgeving. - Willumsen, L. G., et al. (2011). Modelling transport. Wiley-Blackwell. - Wilson, A. (2013). Entropy in urban and regional modelling (routledge revivals). Routledge. - World Health Organization. (2020). Who guidelines on physical activity and sedentary behaviour [Accessed: 2025-05-05]. [https://iris.who.int/bitstream/handle/10665/337001/](https://iris.who.int/bitstream/handle/10665/337001/9789240014886-eng.pdf) [9789240014886-eng.pdf](https://iris.who.int/bitstream/handle/10665/337001/9789240014886-eng.pdf) - Zawadzki, M. J., Smyth, J. M., & Costigan, H. J. (2015). Real-time associations between engaging in leisure and daily health and well-being. Annals of Behavioral Medicine, 49 (4), 605–615. - Zhong, S., & Bian, L. (2023). How regularly do people visit service places? Computers, Environment and Urban Systems, 99, 101896. [https://doi.org/https://doi.org/10.1016/](https://doi.org/https://doi.org/10.1016/j.compenvurbsys.2022.101896) [j.compenvurbsys.2022.101896](https://doi.org/https://doi.org/10.1016/j.compenvurbsys.2022.101896) - Zhou, Z., Chen, A., & Wong, S. (2009). Alternative formulations of a combined trip generation, trip distribution, modal split, and trip assignment model. European Journal of Operational Research, 198 (1), 129–138. [https://doi.org/https://doi.org/10.1016/j.](https://doi.org/https://doi.org/10.1016/j.ejor.2008.07.041) [ejor.2008.07.041](https://doi.org/https://doi.org/10.1016/j.ejor.2008.07.041) # Illustrative example of the accessibility indicator and the whole context of transport poverty. To illustrate the application of the accessibility indicator, the following example outlines the necessary information and steps required to conduct the analysis effectively. #### Case Study: Addressing Transport Poverty in the Province of Toledo. The Province of Toledo seeks to mitigate transport poverty within the region (204 municipalities). To achieve this, they have decided to implement the methodology proposed in this document. Their primary objective is to identify rural areas mostly affected by transport poverty, enabling the efficient allocation of transportation-related financial resources to counter it. Note: this whole example is not based in real data, but serves as a preview of what could be obtained by using it. This example also covers the whole methodology exposed in Figure [3.1,](#page-45-0) however emphasis is put in accessibility analysis, (the target of this work) only briefly covering the other dimensions assessment. 1. Defining the Level of Aggregation: Regional Scope and Population Segmentation. The community establishes the level of aggregation based on both settlement size and population characteristics. If a village has 4,000 or fewer inhabitants, it is considered as a single unit of analysis. For larger settlements, subdivisions are made according to postal codes. In terms of population segmentation, four primary groups are identified: - Young citizens (aged 14 to 18 years) - Working-age population - Retired citizens - Mobility-restricted individuals Recognizing that transport poverty is likely to be more prevalent among lower-income households, the community further categorizes each group into two subgroups based on household income: Households with income > 2,000€ per month, and Households with income < 2,000€ per month. This results in a total of eight distinct population segments. The chosen level of aggregation is based on the availability of current data and aims to ensure a comprehensive and equitable assessment of transport accessibility challenges. 2. Definition of Prototype Citizens for Each Segment. The community authority did not have the means to launch a survey and, due to the lack of detailed household-specific data, opted for a Prototype citizen approach. Therefore, the objective of this section is to define a representative prototype citizen for each segment. This allows for the determination of key characteristics such as gender, place of residence, vehicle ownership (e.g., car, bicycle), any special mobility requirements, and the associated income level. A significant amount of data is required to generate reliable indicators. The community followed these steps for each area: - Gender: Female was always chosen as the representative gender. This decision was particularly relevant for the safety requirements (included in the adequacy dimension), where transportation requirements tend to be stricter for women in order to consider a mean of transport safe. This decision aligns with what is recommended in Cludius et al. [\(2024\)](#page-109-0). - Car Ownership: Using municipal data on the number of cars per household, the municipality applied the following rule: - If the ratio of cars per household in a given municipality was greater than 0.6, both the working-age prototype and the elderly prototype with more than 2,000€/month were considered to own a car. - Otherwise, and for all other groups, no car ownership was assumed. Additionally, the municipality considered including car ownership for young citizens. While individuals aged 14 to 18 cannot drive, their parents often do, and many trips are made using family-owned vehicles. However, as the target group consisted of adolescents, it was ultimately decided that they should be assumed to travel independently. # • Adequacy Requirements: - For the mobility-restricted group, specific mobility requirements were considered. For all other groups, all transports were considered barrier-free. - For safety and information accessibility, due to a lack of available data, transport was assumed to be safe and adequately informative. Ideally, this information would be collected through citizen surveys or interviews with bus drivers. - Income: Due to the absence of household-specific income data at the municipal level, the average income was estimated using the mean income of all citizens in the same province. More precise data was used where available. - Location: To establish representative locations for each income group, two reference points were selected per area, based on the income level. No difference was made according to age, assuming that income level has a greater influence on residence location than age. This approach also helped minimize the amount of required data. Therefore, for each area of analysis, only two points were geocoded: - One location for individuals with household incomes above 2,000 €/month. - One location for individuals with household incomes below 2,000 €/month. An initial request was sent via email to the mayors of all 204 municipalities, asking them to provide two representative addresses. These addresses were subsequently geocoded. For municipalities that did not respond, the centroid of the municipality was chosen as a representative point. 3. Identification of Availability, adequacy and affordability. Based on the previously outlined information, an analysis was conducted for each group to determine the available transportation options. This assessment included GIS calculations to measure accessibility from selected locations to public transport stops (availability), predefined assumptions regarding transport suitability (adequacy), and an evaluation of the proportion of household income spent on public transport (affordability). Additional parameters were included, such as the walking speed of each group and the (in)possibility to use bikes. The following Table [2](#page-118-0) summarizes the results obtained for one of the villages under study. Table 2: Transport availability by income level and population group. Sample Table for one area of study in the illustrative example. | | >
2000
€/month | <
2000€/month | |---------------------|----------------------|------------------| | Young citizens | Bus, bike, train | Bus, bike | | Working age | Car, bus, bike | Bus, bike | | Elderly | Car | Bus | | Mobility restrained | No adequacy | No affordability | 4. Accessibility Analysis. At this point, and only for the means of transport that had proven to be eligible for each group, the accessibility analysis was performed. First, the essential facilities (services) for each group were determined by the government members, as well as the reasonable amount of time for each group to reach each destination, and the frequencies of use of each service. Significant weight was awarded to green areas as the political board was very committed to health and access to nature. Then using OpenStreetMap data, an accessibility analysis was conducted by considering public transport routes, car routes, waiting times, and the locations of essential facilities in Toledo. This process allowed for the completion of Tables [3](#page-120-0) and [4,](#page-121-0) which would be generated for each region under study. To provide a clearer understanding of the results, a sample of what a table could look like for one of the villages under study is presented. Relevance of the example. Although not based in real data, this example reflects well the motivations and thinking behind this analysis. The objective is to provide a standardized framework, generic enough to contemplate the diversity of issues concerning transport poverty, but insightful enough to derive useful and applicable outputs. A proper example of the accessibility indicator is developed in Chapter [4.](#page-62-0) Table 3: Results of accessibility analysis for the example for a village under study, citizens with <2,000 €/month. An illustrative example. | Group | Essential
facilities | Reasonable
time | Number
of trips
per
month | Does it
comply | Accessibility | Number
of
residents | |-----------------------|--------------------------------------------------------|--------------------|------------------------------------|-------------------|---------------|---------------------------| | Young
citizens | Education | 20 | 20 | 1 | | | | <
2,000
€/month | Leisure
area
(cinema,
shopping
mall, etc.) | 45 | 2 | 0 | 37/39 = | 70 | | | Park/Green
Area | 15 | 16 | 1 | 95% | | | | Health
centre | 30 | 1 | 1 | | | | Working
adult | Work | 30 | 20 | 0 | | | | <2,000
€/month | Grocery
shopping | 20 | 4 | 1 | | | | | Sport
centre | 30 | 12 | 1 | 19/39 = | 458 | | | Leisure
area
(cinema,
shopping
mall, etc.) | 45 | 2 | 1 | 49 % | | | | Hospital | 30 | 1 | 1 | | | | Retired | Grocery
shopping | 20 | 10 | 1 | 14/34 = | | | <2,000
€/month | Green areas | 20 | 20 | 0 | 41 % | 573 | | | Hospital | 30 | 4 | 1 | | | | Total | | | | | 47.7% | 1,101 | Table 4: Accessibility results for the population of a village under study. An illustrative example. | Group | >2,000
€/month | <2,000
€/month | | | |------------------------|-------------------|--------------------------|--|--| | Young citizens | 95% (50) | 100% (70) | | | | Working age | 100% (563) | 73% (458) | | | | Elderly | 100% (502) | 41% (573) | | | | Mobility
restrained | No adequacy (10) | No affordability
(12) | | |