Revitalizing Walkability Scores: A New Assessment Based on

Accessibility

J. Bracquart

, T. Leduc

, V. Tourre

and M. Servi

eres

Nantes Universit

e, ENSA Nantes,

Ecole Centrale Nantes, CNRS, AAU-CRENAU, UMR 1563, F-44000 Nantes, France

Keywords:

Walkability, GIS, Accessibility.

Abstract:

Context/Purpose. With the motivation to evaluate the suitability of a urban environment for pedestrian mo-

bility, we revisit walkability scores from the scientiﬁc literature whose one of the most representative ﬁgure is

the Walk Score, which is also commercially exploited through an eponymous website. Methods. This study

reﬁnes the purpose of the foundations of walkability scores as a “pedestrian level of accessibility” score and

mobilizes works on accessibility to simplify the computations of the score. The parameter “importance of an

amenity” gains a more generic estimation method. Lastly, the scoring is proposed for different trip categories

before being aggregated into a global score. Results. We obtain a new scoring, which we apply to a small

town with a simple urban layout for illustration purposes before computing scores for a larger and more di-

verse area. The scores allow us to identify different urban fabrics associated with different opportunities within

walking distance. Conclusions. In the end, we provided walkability scores with a more scalable, explainable

and readable methodology which led to improve their usefulness.

1 INTRODUCTION

1.1 Context

We are experiencing an acceleration of global warm-

ing, the consequences of which are becoming life-

threatening for a growing part of the world’s popula-

tion (GIEC, 2023). Therefore, there is a growing in-

terest in rethinking our current car-oriented mobility

in favor of active forms of mobility such as walking

and cycling in order to meet Sustainable Development

Goal 11 of the United Nations

. Active mobility also

has great beneﬁts for public health (Giles-Corti et al.,

2016), the economy (Santos et al., 2023) and the gen-

eral quality of life (Rosso et al., 2011).

We consider that the will to satisfy a need is the

ﬁrst determinant of the adoption of a mobility, and a

necessary condition for the choice of pedestrian mo-

bility is access to a suitable offer within a “walking

distance”.

https://orcid.org/0009-0000-1633-6447

https://orcid.org/0000-0002-5728-9787

https://orcid.org/0000-0003-4401-9267

https://orcid.org/0000-0001-5749-1590

https://www.un.org/sustainabledevelopment/cities/

(accessed 11/2023)

Here, urbanism plays a key role in the distribution

of modes of transportation for mobility with the idea

that the “best mobility is no mobility”, as expressed

in concepts such as “15-minute cities” (Moreno et al.,

2021), where everyone has access to services they

need to live, within a 15-minute walk or bike ride

from their home. Additional mobility is then used for

leisure and not out of necessity.

1.2 Problematic

This “15-minute cities” concept hence relies on a di-

versity of amenities (places of employment, leisure,

shopping. . . ) accessible (in this article, we under-

stand this term to mean “reachable”) within a “walk-

ing distance” (a distance that people are willing to

walk such as in (Yang and Diez-Roux, 2012)) which

we therefore consider as the basis of a walkable envi-

ronment.

The study of walkable environments is what drives

our research. In the literature, works about walkabil-

ity try mainly to ﬁnd out where people would actually

walk as exposed by the review (Hall and Ram, 2018).

And if the latter answer a key question about the suit-

ability of the urban environment for walking, the im-

pact of their conclusions would beneﬁt from a more

scalable, explainable and readable methodology.

Bracquart, J., Leduc, T., Tourre, V. and Servières, M.

Revitalizing Walkability Scores: A New Assessment Based on Accessibility.

DOI: 10.5220/0012557500003696

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 10th International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2024), pages 39-50

ISBN: 978-989-758-694-1; ISSN: 2184-500X

Elements for this improved methodology can be

found in the ﬁeld of research on “accessibility” to

amenities. Some studies such as (Gastner and New-

man, 2006) or (Xu et al., 2020), for example, attempt

to optimize the positioning and quantity of ameni-

ties in a geographic area to suit as many people as

possible, while other summarize the variety of ser-

vices available to everyone in an “accessibility score”

(Nicoletti et al., 2023).

Therefore, we are motivated to develop a tool to

assess the efﬁciency of the urban layout in terms of

the available amenities so as to promote pedestrian

mobility. In this study, we will answer the follow-

ing question: “How can we evaluate, at the level of a

whole city and with street segments as spatial units,

the degree of pedestrian accessibility achieved by the

Points of Interest (POIs), according to the opportuni-

ties they offer and their location in the urban plan?”.

1.3 Content of this Work

This work aims to revisit the foundations of walkabil-

ity scores by focusing on the degree of accessibility

offered by amenities, assuming that the decision to

make a journey on foot begins with the ability to meet

one’s needs within a walkable distance.

Building on accessibility studies, we will refor-

mulate the purpose of the Walk Score as “Pedestrian

Level of Accessibility Score” and simplify the calcu-

lation by keeping only two parameters: the distance

to the surrounding “amenities” and their relative “im-

portance” within the mix of amenities. This second

parameter, the “importance of an amenity”, will ben-

eﬁt from a more general estimation method. Finally,

we will categorize the trips to calculate a score per

category of trips as a complement to a global walka-

bility score.

2 STATE OF THE ART

2.1 Brief History of the Walk Score

(Hall and Ram, 2018) shows that the Walk Score is

a reference tool to summarize the available amenities

and city layout. From its original publication as the

Walkability Index, what is now known as the Walk

Score has evolved through various iterations, which

we report on here.

2.1.1 Walkability Index

Among the tools to assess the walkability of cities,

“Walkability Index” is a reference developed by

(Frank et al., 2005). It is based on the recognized

framework of the 5Ds (density, diversity, design, dis-

tance to transit and destination accessibility), which

provides a methodology to assess the characteristics

of a built environment that promote walking (Ewing

and Cervero, 2010).

The index is calculated as a combination of four

geographic variables: net residential density, intersec-

tion density, land use mix, and the ratio of retail space

to ﬂoor area (as a later addition).

Each of the variables is weighted based on soci-

ological and demographic studies. Originally devel-

oped to estimate where people actually walk, the sys-

tem has since been widely used and studied (Hall and

Ram, 2018).

2.1.2 Walk Score

This index was later transformed into a “Walk Score”

used on a commercial website

with some variations

to inform the public about the attractiveness of a

neighborhood in terms of available amenities.

(Frank et al., 2021) reports some parts of the

methodology used (some details are protected by

Score is calculated by evaluating the straight-line dis-

tance from a starting point to the nearest destinations

in nine different categories, including grocery stores,

restaurants, general stores, coffee shops, banks, parks,

schools, bookstores, and entertainment venues. For

all categories except restaurants (where ten are con-

sidered), cafes (where two are considered) and stores

(where ﬁve are considered), only the closest destina-

tion is considered. This adjustment takes into account

the increased attractiveness of several nearby comple-

mentary amenities. The distances are then converted

into amenity scores on a scale of 0-100 using a poly-

nomial distance decay function. These amenity scores

are weighted based on research ﬁndings and then ag-

gregated to obtain the ﬁnal walk score, which is also

on a 0-100 scale. Finally, a 0-5% deduction is made

that takes into account the average block length and

intersection density within a 1.5 mile radius of the

starting point.

2.1.3 Street Smart Walk Score

The methodology of the “Walk Score” was later re-

vised into a “Street Smart Walk Score” by the origi-

nal authors of the “Walkability Index” (Frank et al.,

2021), whereby the Walk Score was assessed from an

address primarily using distances based on the street

https://www.walkscore.com/methodology.shtml

(accessed 11/2023)

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

network and the function of decay with distance was

adapted to different categories of amenity.

2.2 Other Works of Interest

2.2.1 Accessibility

Originally, (Hansen, 1959) deﬁned accessibility as

a distance-weighted property of a Point of Interest

(POI). More generally, accessibility measures travel

costs and the quality/quantity of opportunities, as

shown in the review (P

aez et al., 2012). Accessibil-

ity can therefore be measured from a population per-

spective (e.g. the availability of services) or from a

destination perspective (e.g. the catchment area).

(Nicoletti et al., 2023) classiﬁes Points of Interest

into seven categories inspired by Maslow’s hierarchy

of needs (Maslow, 1943). Then, the closest point of

interest in each category is considered to assess ac-

cessibility in each road segment. The ﬁnal score is

calculated as an aggregation of distances, with the im-

portance of the categories being weighted by a panel

of experts.

2.2.2 Studies Regarding Walkability Scores

(Zhao et al., 2021) shows that it is important to con-

sider the pedestrian network instead of the road net-

work in the analysis, especially for Asian cities. Re-

garding the incentive of amenities to walk, the article

sorts the amenities into ten categories (an 11

gives

a “mixing index”) and applies a decay by network-

based distance but also by the number of amenity

units to account for redundancy. In their work, the

authors want to adapt the algorithm for densely popu-

lated Hong Kong.

(Gorrini et al., 2021) evaluates the degree of walk-

ability based on four criteria: accessibility, comfort,

safety and attractiveness, whose values are normal-

ized, distributed in deciles and summarized in a ﬁnal

score.

(Lam et al., 2022) applies the Walk Score al-

gorithm with the additional parameters of “green-

ery” and “proximity to public transport” to the entire

Netherlands and compares the results with those of

a national study on mobility, in which some correla-

tions were found.

2.2.3 About a Walkable Distance

In studies of walkability, the question arises as to the

distance that is to be considered “walkable” and can

serve as a reference for the evaluation of a “degrada-

tion with increasing distance”’ in the convenience of

an amenity. (Yang and Diez-Roux, 2012) reports from

a 2009 survey of 300,000 people in the USA who had

walked a total of more than a million trips, that the

mean and median distances walking were 0.7 and 0.5

miles (or 1.13 and 0.80 kilometers) respectively.

2.3 Comments

2.3.1 Purpose of the Index

As for the original purpose of the index, various ef-

forts have been made to ﬁnd a correlation between

the calculated values and actual pedestrian activity on

the streets, but as noted in the review (Hall and Ram,

2018), the scoring system is best suited to assess the

quality of a built environment in terms of connectiv-

ity and the distribution of amenities. For example,

the tool does not take into account the sociological

and environmental factors that are part of the decision

process to walk.

2.3.2 Choice of Parameters

We assume that the parameters “connectivity” and

“block length” in the Walk Score formula are intended

to capture psychological motivations and deterrents

to walking caused by the urban environment, as de-

scribed by (Lynch, 1960) in “The Image of the City”.

Ultimately, however, these parameters appear to

serve as a calibration of the formula to extrapolate

the calculation of the score from a limited sample of

surveyed people. As a result, the index/score loses

in scalability as it relies on surveys with speciﬁc

and limited spatial coverage, and furthermore it loses

explainability as the formula becomes increasingly

complex with parameters that are not so simple and

meaningful. Furthermore, as already mentioned, the

intended purpose of predicting walkability is not fully

achieved.

This also applies to the idea of adjusting the de-

crease in importance of a POI with increasing dis-

tance by a formula adapted to different categories of

facilities: to reﬁne the index/score so that it matches

the observed walks, it again loses scalability and ex-

plainability.

2.3.3 Importance of the Amenities

The Walk Score and other studies such as (Su et al.,

2017) rely on surveys to assess the importance of

amenities, which has several limitations, it is not ex-

haustive in terms of the variety of amenities, it re-

quires some resources to conduct the survey at a suf-

ﬁcient scale and it does not necessarily translate well

to other populations and locations (limitations in scal-

ability mentioned earlier).

Revitalizing Walkability Scores: A New Assessment Based on Accessibility

2.3.4 Exposing the Results

The various studies that use the Walk Score such as

(Gorrini et al., 2021) and (Lam et al., 2022) or one of

its variations, often end up with an index/score ren-

dered on surfaces, while we want to evaluate the walk-

ability of paths. With the move to network-based dis-

tances to amenities, it even seems like a missed oppor-

tunity to not render the scores directly on the paths.

Finally, this reduces the readability of the index.

2.3.5 Accessibility

The works on accessibility mobilize intuitive and

therefore more explainable methods. They must be

tailored to the study of walkability and take into ac-

count the largest possible range of amenities that

speciﬁcally induce pedestrian mobility. Estimation of

the importance of amenities also exhibits signiﬁcant

variations between studies, usually with limited cov-

erage on the diversity of amenities.

3 PROPOSED INDEX

3.1 Purpose of the Score

First, we propose clarifying the purpose of the index

by pursuing the objective of evaluating “pedestrian

accessibility”, which reﬂects the efﬁciency of the ur-

ban layout. We do not intend to explain the observed

walking behavior. From the point of view of the ur-

ban planner, the question answered by the scoring can

be formulated as follows: “how well equipped is a

particular street in terms of facilities within walking

distance and their respective distance”. From the cit-

izens’ point of view, it gives a “convenience of living

somewhere” independently of an external means of

transportation. The point here is to ask a more direct,

easy-to-answer question. So we hope to gain in ex-

plainability. In the following, we will refer to our

score as “pedestrian level of accessibility score” in-

stead of “walkability score” to reﬂect this adaptation

of purpose.

Our index is computed in respect of the following hy-

potheses:

• we consider statistically averaged persons in

terms of physical abilities and needs

• all street segments are of the same perceived qual-

ity (street width, vegetation, etc.) and the altime-

try is not considered

• all amenities in a category are of the same quality

• we consider mobilities at daytime on an averaged

day

3.2 Simpliﬁcation of the Score

Calculation

Our walkability index has its roots in the approach

used in accessibility studies such as those conducted

by (P

aez et al., 2012). To align with this approach,

we simplify the calculation of the walkability score

keeping only two parameters: a decay with increasing

distance to amenities and their relative importance in

the mix of amenities.

Score =

amenities

∑

k=0

Decay(dist

) × Importance

(1)

This decay function is taken strictly linear (see the

green line on ﬁgure 1) and not speciﬁc to a given cat-

egory of amenity to gain in explainability. Addition-

ally, in our study, the distance itself is evaluated using

the network of streets as in the latest iteration from

(Frank et al., 2021) which improves precision in com-

parison to using distances along straight lines.

Decay(dist) = 1.0 −

dist

walkable dist

(2)

if dist ≤ walkable dist, else Decay(dist) = 0.0

Figure 1: Cumulative percentages of walking trips by dis-

tance, red dots are from a survey, blue line is an interpola-

tion and green line is our simpliﬁcation as a linear regres-

sion which captures about 80% of the cumulative percent-

ages of walking trips by distance (adapted from (Yang and

Diez-Roux, 2012)).

3.3 New Method to Estimate the

Parameter “Importance of an

Amenity”

Every day, all the people in an area set out to satisfy

their needs. Given a sufﬁciently large and contiguous

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

geographic urban area, we can even approximately as-

sume that all amenities satisfy all needs of the popula-

tion (daily trips in and out of the area become anecdo-

tal). We can therefore assume that an amenity satisﬁes

the needs of the population on average up to its daily

attendance (as a number of persons) and we propose

to consider this as the importance of the amenity.

It should be noted that considering a large area

also allows for statistical averaging and helps to elim-

inate a number of considerations relating to the in-

habitants on the one hand (age, gender, socio-cultural

background, individual sensitivities, skills. . . ) and the

amenities on the other hand (quality, individual eco-

nomic issues, particular type of business. . . ).

This method of evaluation has the advantage of

being universal (all amenities are visited) and objec-

tive (it can be measured). We know that this data can

be found (Google Maps offers hourly attendance as a

tip, e.g. “more or less frequented than usual”). How-

ever, for our study we were unable to gather sufﬁcient

data, so we initially estimated this parameter of daily

attendance using assumptions. In terms of a com-

mercial application, there would be a simple method

to accurately determine the importance of each POI

without relying on a survey that is limited in its geo-

graphical coverage and in the variety of facilities con-

sidered.

3.4 Computation of Scores per Trip

Category and Their Synthesis

To make our index more useful, we compute scores

for different trip categories deﬁned in the general sur-

vey on mobilities (INSEE, 2010) to which we can as-

sociate various amenities.

Trips categories and their related amenities:

• “groceries and shopping”: shopping centers,

malls, downtown shops. . .

• “work”: places of employment

• “leisure”: sport centers, restaurants, coffee shops,

cultural places, entertainment venues. . .

• “studies”: any educational establishment from el-

ementary school to universities

• “administrative and healthcare”: public adminis-

trations, hospitals. . .

• “visits”: friends and parents to visit

Relying on open databases, we managed to gather

a quite exhaustive list of the geolocalized amenities to

represent the trip categories (see data sources in sec-

tion 4.4.1). We interpreted the category “visits” as

the potential for a social life proportional to the pop-

ulation localized at a walking distance (as we con-

sider averaged persons in an averaged city). Looking

for the number of inhabitants per street, we catego-

rize the building footprints according to the land use

(see ﬁgure 2) and associated a number of inhabitants

to each: 3 for “habitations low density”, and a value

proportional to the building’s height and its ﬂoor area

for “habitations medium density” or “city center”. In

the end, we make the assumption that each building

becomes an attractor in proportion to the number of

inhabitants occupying it.

The scores per trip category are then combined by

a weighted sum to a “global score” using each cat-

egory’s share in the repartition of trips (see table 1)

as the weighting that we ﬁnd in the aforementioned

survey (INSEE, 2010).

Score =

categories

∑

k=0

category score

× weight

(3)

3.5 Exposing the Results

In order to improve both readability and explainabil-

ity, the not null scores computed for our large geo-

graphic area were divided into quintiles for each trip

category (e.g. the 1/5 of the highest scores in trip cat-

egory “work” on all road sections form the ﬁrst quin-

tile for this category). This distribution can then be

used as a normalization for the scores in different ge-

ographical areas, allowing comparative analysis.

The resulting scores are rendered on the network

of streets whose walkability is being estimated. This

should improve readability.

It is also informative to display a score per cate-

gory of trips as well as the combined score described

in 3.4.

It should be noted that the color scale is adapted

from the European energy label for household appli-

ances, which is already familiar to the general public.

4 IMPLEMENTATION

4.1 Computational Methods

Inspired by methods from numerical simulations we

turn the equation (1) used to evaluate the scores for a

trip category, into the following computation:

S = A × b (4)

with A a square matrix of the decays in im-

portance related to the distances from every streets

segments to every streets segments, S the unknown

scores and b the vector of the importance of the

Revitalizing Walkability Scores: A New Assessment Based on Accessibility

amenities for a trip category associated to their clos-

est street segment. A is very sparse as decay is zero

above the reference walkable distance between two

street segments.

This way, the computational cost relies almost en-

tirely on the construction of the matrix A and allows

to evaluate the level of opportunity offered by almost

any number of amenities to almost no further compu-

tational cost.

4.2 Algorithmic Translation of the

Index

The main steps for the computation of the score are:

• conversion from a pedestrian network of streets

describing the city to a topology of streets seg-

ments (list of neighbors)

• computation of the matrix A reporting the decay

with distance coefﬁcients using equation (2) and

taking the network based distance between each

pair of streets considered

• each amenity is classiﬁed in a trip category

• each amenity is given an importance (see 3.3)

• projection of the amenities onto the geometrically

closest street segments (to construct vector b)

• computation of the product S = A × b (equation

(4)) for each trip category

• normalization of the scores per quintiles against

the scores at a metropolitan scale

• rendering of the score on the streets network for

each trip category

• combination of the scores to get one ﬁnal “pedes-

trian level of accessibility” score

• rendering of the ﬁnal score on the streets network

4.3 Softwares Involved

The method has been implemented in Python code

with GIS algorithms from GeoPandas (Jordahl et al.,

2020). The construction of matrix A has been thor-

oughly optimized by leveraging SciPy’s (Virtanen

et al., 2020) sparse matrices and NumPy’s (Harris

et al., 2020) high-performance C-based operations.

The renders are performed with Folium (python-

visualization, 2020).

4.4 Data Sources

4.4.1 Geographic Data

The street network used in the study (and curated

to keep only the pedestrian network) originates from

the IGN

(French National Institute of Geographic

and Forest Information). The amenities are ex-

tracted from Open Street Map

for categories “gro-

ceries and shopping”, “leisure”, “studies”, “admin-

istrative/healthcare”, from the SIRENE database

for

category “work” and from the IGN for category “vis-

its”.

4.4.2 Sociologic Data

The sociologic elements from this study comes from

the INSEE

(National Institute of Statistics and Eco-

nomic Studies) which is a public institution attached

to the French Minister of the Economy, Finance and

Industrial and Digital Sovereignty. Its mission is to

collect, analyze, and disseminate information about

the French economy and society throughout its entire

territory.

5 RESULTS

5.1 Description of the Case Study

We will ﬁrst illustrate the proposed scoring system

computing scores for the town of Sautron

(French

city with 8,473 inhabitants in 2020) which has a rela-

tively simple urban layout.

The city is organized around a central street, a

small city center and some sports facilities on the

western side. The town hall of Nantes is located to the

southeast, about eleven kilometers away. Sautron is

surrounded by an industrial area (to the east), a high-

way to the south and some vegetation (both agricul-

tural and natural) to the north and west sides.

We will then apply our scoring on the vaster

and more heterogeneous metropolitan area of Nantes

etropole (French city of Nantes with suburbs,

665,204 inhabitants in 2019

Nantes M

etropole consists of a historic center lo-

cated in the middle of a ring road (see ﬁgure 2). It

is divided horizontally by the Loire River, which cuts

out a large island below the city center. The area is

divided into twenty-four municipalities, with Nantes

being the most important.

https://geoservices.ign.fr/bdtopo (edition 2023-06-15)

http://download.openstreetmap.fr/ (accessed 11/2023)

https://www.sirene.fr (accessed 11/2023)

https://www.insee.fr (accessed 11/2023)

https://fr.wikipedia.org/wiki/Sautron (accessed

11/2023)

https://fr.wikipedia.org/wiki/Nantes Metropole (ac-

cessed 11/2023)

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

Figure 2: General map of Nantes M

etropole and location of

Sautron in the northwest.

We take Nantes M

etropole as our “large area of

reference”.

The distribution of trips per purpose discussed in

section 3.4 is taken from the generic survey on mobil-

ity (INSEE, 2010):

Table 1: Distribution of trips according to their purpose for

Nantes M

etropole (INSEE, 2010).

Purpose Weekday Saturday Sunday

Groceries and shopping 20% 35% 10%

Work regular place 19% 4% 1%

Leisure 16% 24% 23%

Studies 12% 4% 0%

Visits 10% 17% 12%

Administrative and healthcare 6% 3% 4%

Other 17% 13% 50%

Then, the weight for trip purpose k and for an av-

eraged day is:

× w

weekday

× (w

saturday

+ w

sunday

) (5)

These statistics are obtained with the following

methodology:

• a trip is deﬁned as a displacement from a destina-

tion to another, a new trip is registered each time

a purpose is achieved

• return trips are not considered

• respondents are six years old and above

Regarding the reference walkable distance (see 2.2.3),

for our investigation, we consider a value of 1.0 miles

(1.6 kilometers) for the whole round trip according

to the green line on the ﬁgure 1: at 400 meters the

importance of an amenity is divided by two and at

800 meters (1600/2) it becomes null.

5.2 Results for Sautron

5.2.1 Scores per Trip Categories

We apply our algorithm to the city of Sautron and

obtain the values for pedestrian accessibility scores

showed in the ﬁgure 3.

(a) Work: we see that most of the work offers comes

from the industrial zone on the east as well as the in-

stitutions of the city center and the shops on the main

pathway

(b) Shopping / Groceries: most of the shopping of-

fers is situated near the central road

entertainment venue at the west of the city center

(d) Health / Institutions: most of the institutions take

place around the city hall at the city center

(e) Studies: the schools are mainly located at the city

center in Sautron

(f) Visits: most of the population is concentrated

around the central road and city center with the den-

sity getting lower further away

Satellite imagery as well as local cartography al-

ready provided this kind of information, so the added

value here comes from the synthesis as well as the dis-

play of the geographic scope of the service provided

by the amenities.

We also gain a comparison with Nantes M

etropole

against which the scores were normalized. In most

categories, the scores experience a quick decay with

the increasing distance as the amenities are in a low

number and quite concentrated.

5.2.2 Global Score

We combine the scores from the different categories

into one global “pedestrian level of accessibility

score” using, as weights, the distribution of trips from

(INSEE, 2010) described previously (see 5.1). We get

ﬁgure 4.

For Sautron, as the best score “A” is essentially

absent, we can say that its city center is not as

well provided with amenities as the center of Nantes

etropole and that’s why the city beneﬁts from good

public transportation to the latter. People living in

streets ranking “C” can carry out some activities on

foot but will probably appreciate using a bicycle to

gain better access to the whole mix of Sautron’s

amenities. In comparison, the population on the

streets ranking “D” and “E” have an increased de-

pendency on motorized vehicles (which will be con-

ﬁrmed analyzing Nantes M

etropole in the next para-

graph, these two scores being quite bad).

5.3 Results for Nantes M

etropole

Nantes M

etropole is a more heterogeneous territory

than Sautron, composed of a variety of urban environ-

ments ranging from a dense historic center to small

communities and interurban places. Nature, with its

Revitalizing Walkability Scores: A New Assessment Based on Accessibility

Figure 3: Scores and amenities (size related to importance) associated to work (a), shopping (b), leisure (c),

health/administration (d), studies (e) and visits (f) normalized over the results on Nantes M

etropole.

Figure 4: Global pedestrian level of accessibility score for

Sautron.

parks and rivers, and industrial areas also characterize

the area.

We evaluate the scores from the different cate-

gories of amenities as well as the overall rating on

Nantes M

etropole. For the sake of brevity, we only

report on the latter here (see ﬁgure 5).

Rank A: the city center 1 unsurprisingly yields ex-

cellent scores ranking A in every trip categories with

amenities in over abundance to serve also the neigh-

boring districts and municipalities. With 2 , 3 and

4 we identify secondary centers, and most notably

their geographic extent, which are also well equipped

with amenities in all categories. Also, 4 encom-

passes the biggest building of Nantes M

etropole (Sil-

lon de Bretagne, about 2,500 inhabitants) which is lo-

cated close to an industrial busy road/area with jobs

and shopping offerings. It should be noted that the

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

Figure 5: Zoom on the scores on the western districts of

Nantes M

etropole obtained for a round trip walkable dis-

tance of 1.6 kilometers.

good ranking fails to report that the area is quite un-

pleasant to reach on foot, as it is mainly designed for

car-oriented mobility.

Rank B: 5 has slightly fewer offers in shopping and

employment than the city centers, 6 is an indus-

trial area that is hardly inhabited but offers employ-

ments and shopping opportunities for people willing

to come with motorized means of transport. Last with

this rank, 7 only maintains a good score due to its

proximity to two large and well-provided avenues.

Rank C: 8 is a residential area which gets a good

employment rating from the industrial zone nearby,

however it quite fails to answer the needs of its in-

habitants in all the other categories. A good opening

towards the better equipped east prevents it from get-

ting a worse notation.

Rank D: as with 8 , the place 9 lacks service in

most amenity categories and is further convoluted, re-

ducing access to better equipped places.

Rank E: 10 offers a place with poor access to any

amenities and additionally with rather convoluted ac-

cess. Its inhabitants are probably quite dependent on

cars for their mobility.

5.4 Sensitivity Analysis of the Decay

Function

The decay function described in section 3.2, reports

on the decrease in opportunities offered by an amenity

as the distance to it increases.

It depends on a walkable distance (see 2.2.3) that

we chose at 1.6 kilometers (total round trip) for our

case study (see 5.1). In order to characterize the sensi-

tivity of this walkable distance, we compute our scor-

ing with a value decreased by 50% to 0.8 kilometers

and increased by 100% to 3.2 kilometers.

Figure 6: Zoom on the scores on the western districts of

Nantes M

etropole obtained for a round trip walkable dis-

tance of 0.8 kilometers (a) and 3.2 kilometers (b).

The result for the western districts of Nantes

etropole is exposed on ﬁgure 6 to be compared with

ﬁgure 5.

Regarding the decreased walkable distance (ﬁg-

ure 6.a) we observe that only the places well furnished

with amenities keep a good score, while those that

mainly beneﬁted by well provided nearby places get

well degraded scores. As such, the residential only

districts 7 , 8 and 9 have their scores devalued,

while well-provided environments get some credits.

Indeed, with this analyze, the local environment gains

a greater importance and we can say that this choice

of walkable distance would better reﬂect the mobil-

ity of senior people who are known to walk smaller

distances.

About the increased walkable distance (ﬁgure 6.b)

most of the places considered get an improved score

compared with round trip walkable distance of 1.6

kilometers as they beneﬁt from the proximity with the

Revitalizing Walkability Scores: A New Assessment Based on Accessibility

Figure 7: Scores on various French cities: Lille (a), Rennes (b), Paris (c), Strasbourg (d), Bordeaux (e) and Grenoble (f).

center of Nantes M

etropole. However, some places,

such as 3 get a degraded score. This is because the

normalization process described in section 3.5 leads

to an equal number of streets (with not null scores) in

each rank for the whole Nantes M

etropole.

5.5 Scalability

To demonstrate good scalability of the method, we

compute the scores for six other French metropolitan

areas on the ﬁgure 7.

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

In this ﬁgure, the distribution of scores into quin-

tiles is done independently for each city, which pre-

vents the comparison of absolute scores. However,

we could select one city against which the others are

normalized to allow such a comparison.

A similar ﬁgure could be generated to display the

scores for each category with the associated ameni-

ties. Also, we provide the map to the general public

through a website

6 DISCUSSION AND

CONCLUSION

In our study dedicated to walkability scores, we de-

ﬁned a more direct and easy-to-answer question than

the usual literature (Hall and Ram, 2018) to gain

in explainability. With the same prospect, we also

chose to consider statistically averaged persons in an

averaged city built of averaged streets and amenities.

Revisiting these hypotheses make for as many oppor-

tunities of future studies. Again for explainability,

we simpliﬁed the score calculation taking inspiration

from works on accessibility to amenities. Also, for

scalability, we revised the method to estimate the pa-

rameter “importance of an amenity” to a more uni-

versal and objective approach distinct from the ap-

proaches in the literature which are mainly based on

surveys (Su et al., 2017). The surveys have the ad-

vantage of being quite precise in their conclusions but

they are also quite limited in their geographic cov-

erage and in the diversity of amenities considered.

Unfortunately, with our limited time and resources,

we ended up estimating the parameter with assump-

tions, however as mentioned, the data exists and it

should be possible to acquire it for a commercial ap-

plication. The statistical and geographic data for our

case study otherwise come from very generic sources

which should allow to easily apply the scoring to at

least the whole French territory. We evaluated scores

for different trip categories (for usefulness) and per-

formed their combination into a ﬁnal score. The ren-

dering of the computed scores was also worked out

for readability which offers ﬁner details compared to

scores aggregated on surfaces in (Gorrini et al., 2021)

or (Lam et al., 2022). In the end, we discussed here

only one aspect of the question “where do people ac-

tually walk” behind walkability scores. Our purpose

was indeed to lay a more robust foundation to those

scores, considering that the accessibility to amenities

is the starting point of any mobility.

Additionally, for future developments, we are go-

https://villes-marchables.huma-num.fr

ing to interview urban planners so as to reﬁne the use-

fulness of our scoring and help them make more in-

formed decisions.

ACKNOWLEDGEMENTS

The authors thanks the French Ministry of Higher Ed-

ucation and Research for the ﬁnancial support of this

project.

REFERENCES

Ewing, R. and Cervero, R. (2010). Travel and the Built

Environment: A Meta-Analysis. Journal of the Amer-

ican Planning Association, 76(3):265–294.

Frank, L. D., Appleyard, B. S., Ulmer, J. M., Chapman,

J. E., and Fox, E. H. (2021). Comparing walkability

methods: Creation of street smart walk score and ef-

ﬁcacy of a code-based 3D walkability index. Journal

of Transport & Health, 21:101005.

Frank, L. D., Schmid, T. L., Sallis, J. F., Chapman, J.,

and Saelens, B. E. (2005). Linking objectively mea-

sured physical activity with objectively measured ur-

ban form: Findings from SMARTRAQ. American

Journal of Preventive Medicine, 28(2):117–125.

Gastner, M. and Newman, M. (2006). Optimal design of

spatial distribution networks. Physical review. E, Sta-

tistical, nonlinear, and soft matter physics, 74:016117.

GIEC (2023). Synthesis report of the IPCC sixth assessment

report (AR6).

Giles-Corti, B., Vernez-Moudon, A., Reis, R., Turrell, G.,

Dannenberg, A. L., Badland, H., Foster, S., Lowe, M.,

Sallis, J. F., Stevenson, M., and Owen, N. (2016). City

planning and population health: A global challenge.

Lancet (London, England), 388(10062):2912–2924.

Gorrini, A., Presicce, D., Choubassi, R., and Sener, I. N.

(2021). Assessing the Level of Walkability for Women

Using GIS and Location-based Open Data: The Case

of New York City. Findings, (2019).

Hall, C. M. and Ram, Y. (2018). Walk score® and its po-

tential contribution to the study of active transport and

walkability: A critical and systematic review. Trans-

portation Research Part D: Transport and Environ-

ment, 61:310–324.

Hansen, W. G. (1959). How Accessibility Shapes Land

Use. Journal of the American Institute of Planners,

25(2):73–76.

Harris, C. R., Millman, K. J., Van Der Walt, S. J., Gom-

mers, R., Virtanen, P., Cournapeau, D., Wieser, E.,

Taylor, J., Berg, S., Smith, N. J., Kern, R., Picus, M.,

Hoyer, S., Van Kerkwijk, M. H., Brett, M., Haldane,

A., Del R

ıo, J. F., Wiebe, M., Peterson, P., G

erard-

Marchant, P., Sheppard, K., Reddy, T., Weckesser,

W., Abbasi, H., Gohlke, C., and Oliphant, T. E.

(2020). Array programming with NumPy. Nature,

585(7825):357–362.

Revitalizing Walkability Scores: A New Assessment Based on Accessibility

INSEE (2010). Les transports et deplacements des habitants

des Pays de la Loire.

Jordahl, K., Bossche, J. V. D., Fleischmann, M., Wasser-

man, J., McBride, J., Gerard, J., Tratner, J., Perry, M.,

Badaracco, A. G., Farmer, C., Hjelle, G. A., Snow,

A. D., Cochran, M., Gillies, S., Culbertson, L., Bartos,

M., Eubank, N., Maxalbert, Bilogur, A., Rey, S., Ren,

C., Arribas-Bel, D., Wasser, L., Wolf, L. J., Journois,

M., Wilson, J., Greenhall, A., Holdgraf, C., Filipe, and

Leblanc, F. (2020). Geopandas/geopandas: V0.8.1.

Zenodo.

Lam, T. M., Wang, Z., Vaartjes, I., Karssenberg, D., Ettema,

D., Helbich, M., Timmermans, E. J., Frank, L. D., den

Braver, N. R., Wagtendonk, A. J., Beulens, J. W. J.,

and Lakerveld, J. (2022). Development of an ob-

jectively measured walkability index for the Nether-

lands. International Journal of Behavioral Nutrition

and Physical Activity, 19(1):50.

Lynch, K. (1960). The Image of the City. Publication of

the Joint Center for Urban Studies. M.I.T. Press, Cam-

bridge, Mass., 33. print edition.

Maslow, A. H. (1943). A theory of human motivation. Psy-

chological Review, 50(4):370–396.

Moreno, C., Allam, Z., Chabaud, D., Gall, C., and Prat-

long, F. (2021). Introducing the “15-minute city”:

Sustainability, resilience and place identity in future

post-pandemic cities. Smart Cities, 4(1):93–111.

Nicoletti, L., Sirenko, M., and Verma, T. (2023). Disadvan-

taged communities have lower access to urban infras-

tructure. Environment and Planning B: Urban Analyt-

ics and City Science, 50(3):831–849.

aez, A., Scott, D. M., and Morency, C. (2012). Measuring

accessibility: Positive and normative implementations

of various accessibility indicators. Journal of Trans-

port Geography, 25:141–153.

python-visualization (2020). Folium, https://python-

visualization.github.io/folium/.

Rosso, A. L., Auchincloss, A. H., and Michael, Y. L.

(2011). The urban built environment and mobility in

older adults: A comprehensive review. Journal of Ag-

ing Research, 2011:816106.

Santos, A. C., Willumsen, J., Meheus, F., Ilbawi, A., and

Bull, F. C. (2023). The cost of inaction on physical in-

activity to public health-care systems: A population-

attributable fraction analysis. The Lancet Global

Health, 11(1):e32–e39.

Su, S., Pi, J., Xie, H., Cai, Z., and Weng, M. (2017). Com-

munity deprivation, walkability, and public health:

Highlighting the social inequalities in land use plan-

ning for health promotion. Land Use Policy, 67:315–

326.

Virtanen, P., Gommers, R., Oliphant, T. E., Haberland, M.,

Reddy, T., Cournapeau, D., Burovski, E., Peterson, P.,

Weckesser, W., Bright, J., Van Der Walt, S. J., Brett,

M., Wilson, J., Millman, K. J., Mayorov, N., Nel-

son, A. R. J., Jones, E., Kern, R., Larson, E., Carey,

C. J., Polat,

I., Feng, Y., Moore, E. W., VanderPlas,

J., Laxalde, D., Perktold, J., Cimrman, R., Henrik-

sen, I., Quintero, E. A., Harris, C. R., Archibald,

A. M., Ribeiro, A. H., Pedregosa, F., Van Mulbregt,

P., SciPy 1.0 Contributors, Vijaykumar, A., Bardelli,

A. P., Rothberg, A., Hilboll, A., Kloeckner, A., Sco-

patz, A., Lee, A., Rokem, A., Woods, C. N., Ful-

ton, C., Masson, C., H

aggstr

om, C., Fitzgerald, C.,

Nicholson, D. A., Hagen, D. R., Pasechnik, D. V.,

Olivetti, E., Martin, E., Wieser, E., Silva, F., Lenders,

F., Wilhelm, F., Young, G., Price, G. A., Ingold, G.-

L., Allen, G. E., Lee, G. R., Audren, H., Probst, I.,

Dietrich, J. P., Silterra, J., Webber, J. T., Slavi

c, J.,

Nothman, J., Buchner, J., Kulick, J., Sch

onberger,

J. L., De Miranda Cardoso, J. V., Reimer, J., Har-

rington, J., Rodr

ıguez, J. L. C., Nunez-Iglesias, J.,

Kuczynski, J., Tritz, K., Thoma, M., Newville, M.,

ummerer, M., Bolingbroke, M., Tartre, M., Pak, M.,

Smith, N. J., Nowaczyk, N., Shebanov, N., Pavlyk,

O., Brodtkorb, P. A., Lee, P., McGibbon, R. T., Feld-

bauer, R., Lewis, S., Tygier, S., Sievert, S., Vigna, S.,

Peterson, S., More, S., Pudlik, T., Oshima, T., Pin-

gel, T. J., Robitaille, T. P., Spura, T., Jones, T. R.,

Cera, T., Leslie, T., Zito, T., Krauss, T., Upadhyay,

U., Halchenko, Y. O., and V

azquez-Baeza, Y. (2020).

SciPy 1.0: Fundamental algorithms for scientiﬁc com-

puting in Python. Nature Methods, 17(3):261–272.

Xu, Y., Olmos, L. E., Abbar, S., and Gonz

alez, M. C.

(2020). Deconstructing laws of accessibility and

facility distribution in cities. Science Advances,

6(37):eabb4112.

Yang, Y. and Diez-Roux, A. V. (2012). Walking Distance

by Trip Purpose and Population Subgroups. American

Journal of Preventive Medicine, 43(1):11–19.

Zhao, J., Sun, G., and Webster, C. (2021). Walkability scor-

ing: Why and how does a three-dimensional pedes-

trian network matter? Environment and Planning B:

Urban Analytics and City Science, 48(8):2418–2435.

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management