Research in Transportation Economics 98 (2023) 101263

Contents lists available at ScienceDirect

Research in Transportation Economics

journal homepage: www.elsevier.com/locate/retrec

Research paper

Collective and individual spatial equity measure in public transit

accessibility based on generalized travel cost
Wu Li a, c, **, Haotian Guan b, c, Wenwen Qin b, c, Xiaofeng Ji b, c, *
a
School of Transportation and Logistics, Faculty of Infrastructure Engineering, Dalian University of Technology, Dalian, 116024, China
b
Faculty of Transportation Engineering, Kunming University of Science and Technology, Kunming, Yunnan, 650500, China
c
Yunnan Modern Logistics Engineering Research Center, Kunming University of Science and Technology, Kunming, 650500, China

A R T I C L E I N F O A B S T R A C T

Keywords: This study devises a novel two-dimensional analysis framework for spatial equity in public transit accessibility.
Public transit We incorporated door-to-door travel time and ticket price by public transit into a generalized travel cost function
Spatial equity to measure utility-based accessibility following a log-sum formulation. Then, this study designed a Palma ratio of
Utility-based accessibility
accessibility and a neighboring accessibility gap index to respectively examine the collective spatial equity and
Palma ratio
individual spatial equity of public transit. Finally, we took an empirical case of Kunming city, China to analyze
Neighboring accessibility gap
Door-to-door time the unity-based accessibility distribution and checked collective and individual spatial equity levels reaching
business centers, transport hubs, and 3-A hospitals. The results show that the utility-based accessibility reaching
transport hubs by public transit is less than 50% to business centers and 32% to 3-A hospitals. Collectively, the
expected minimum generalized travel cost by public transit in the poorest 40% of traffic analysis zones is almost
2.5 times as much as in the richest 10% of traffic analysis zones, and the collective spatial equity in old blocks is
superior to new blocks. Individually, the number of traffic analysis zones with no spatial inequity, slight inequity,
and medium inequity in public transit accessibility respectively occupied almost 57%, 38%, and 5%. We also
found individual spatial equity may be associated with the walking distance during taking public transit and the
number of neighboring traffic analysis zones.

1. Introduction 2019).
In the last decade, researchers have devoted significant attention to
Promoting equitable development of public transport system in the exploring the concepts and measurements of equity in public transit.
aspect of social equity (Behbahani et al., 2018; Grengs, 2013) and spatial However, it is more difficult to measure transit equity than to under­
equity (Delmelle & Casas, 2012; Kelobonye et al., 2019) has become a stand one since the principle of equity remains sophisticated and
primary agenda for traffic planners and city engineers. Especially, public diverse. There are three basic categories of transit equity: horizontal
transit has been an essential role to provide necessary access to social equity, vertical equity, and spatial equity in the existing literature. Lit­
activities for low-income people without cars for a long time (Ignaccolo man (2012) developed horizontal equity and vertical equity with in­
et al., 2016; Nahmias-Biran et al., 2013; Stanley & Lucas, 2008). Many come and social class, and Ricciardi et al. (2015) discussed vertical
cities have been aware of the impact of transportation projects (Nah­ equity with mobility demand and ability. Meanwhile, several studies
mias-Biran et al., 2013; Niehaus et al., 2016) and transportation plans attempted to apply the gap of accessibility to identify the difference in
(Behbahani et al., 2018; Camporeale et al., 2016) on equity in different transportation benefits distributed in spatial units, which was usually
aspects, and they attempt to advocate the equity-oriented transport called spatial equity (Kelobonye et al., 2019; Martens et al., 2012). As far
development for enhancing public transport services (R. J. Lee et al., as assessment methods of transportation equity are concerned, their
2016; Oswald Beiler & Mohammed, 2016). Therefore, planning an principles and measures both remain unclear. There is no doubt that
equitable public transport system has potential benefits for the sus­ these proposed methods based on different equitable principles may
tainable and inclusive development of urban mobility (Wang et al., exert some bias when evaluating horizontal equity, vertical equity, and

* Corresponding author. No.727 Jingmingsouth Road, Chengong District, Kunming city, Yunnan Province, China.
** Corresponding author. No.2 Linggong Road, Ganjingzi District, Dalian City, Liaoning Province, China.
E-mail addresses: Liwu625@126.com (W. Li), haotian0813@yeah.net (H. Guan), qinww@kust.edu.cn (W. Qin), yiluxinshi@sina.com (X. Ji).

https://doi.org/10.1016/j.retrec.2023.101263
Received 19 May 2021; Received in revised form 7 February 2023; Accepted 16 February 2023
Available online 24 February 2023
0739-8859/© 2023 Published by Elsevier Ltd.
W. Li et al. Research in Transportation Economics 98 (2023) 101263

others (Golub & Martens, 2014; Martens et al., 2012). accessibility-based equity related to workplaces, dining, and healthcare
In previous works, accessibility has been a prevalent tool to examine has been usually checked.
the spatial equity level in public transit accessibility. The most common Although extensive research has been carried out on public transit
research paradigm concerning accessibility-based equity measures is as accessibility-based equity formulations, to the best of our knowledge,
follows. To begin with, the travel time by public transit will be estimated there is limited research that investigates spatial equity levels in public
as accurately as possible based on a Geographic Information System transit accessibility both collectively and individually in a multimodal
(Tahmasbi & Haghshenas, 2019; Yang et al., 2017). Then, a mathe­ transit network. Therefore, this study is motivated to address the iden­
matical model (such as the gravity model, utility model, space-time tified research niche, and the following contributions are made.
model, and opportunity model) is developed to compute public transit
accessibility considering main trip purposes (Ben-Elia & Benenson, (1) We introduce the generalized travel cost into a utility-based
2019; Hansen, 2007; Kelobonye et al., 2019; L. Li et al., 2017; Zhou accessibility measure that integrates door-to-door travel time
et al., 2018). Finally, the Lorenz curve and Gini coefficient are used to (including walking time, waiting time, transfer time, and travel
measure equity level (Lucas et al., 2015; Ricciardi et al., 2015; Welch, time) and ticket price by public transit. Recalling Table 1 has
2013). Table 1 summarizes related studies on public transit pointed out that accurate travel time plays a key role in public
accessibility-based equity measures in the last four years. In these transit accessibility measures. The difference and novelties of this
studies, transit modes involved common bus and metro modes. study compared with existing literature related to travel time
Recent developments in public transit accessibility-based equity estimation by public transit are summarised in Table 2. On the
measures concentrated on the following four aspects. First, in the aspect one hand, a part of previous studies adopted the self-built transit
of public transit accessibility measures, travel time has become the most network to estimate travel time by public transit. Under this
prevalent indicator for accessibility formulation. Most researchers are workflow, passengers’ walking time and waiting time usually are
used to developing a direct function or composite function of travel time assumed to be constant or directly overlooked. Meanwhile,
to measure public transit accessibility. Second, main three types of congestion effects of background traffic in the road network do
concerning equity measure indicators have been widely used, including not consider in their formulations. On the other hand, previous
several economical indexes (such as the Theil index, Paloma ratio, studies firsthand took travel time rather than generalized travel
Lorenz curves, Gini index, and average consumer surplus), statistical costs to model public transit accessibility, which led to the impact
indicators (such as Getis-Ord statistic and local indicators of spatial as­ of the ticket price on public transit accessibility is not considered.
sociation) and self-proposed indicators (such as accessibility sufficiency (2) We devise a two-dimensional measure framework of spatial eq­
index, travel time gap/ratio). Third, in terms of equity types, most uity in public transit accessibility, including collective spatial
existing studies have been mainly focused on the collective equity level. equity and individual spatial equity. Concerning collective spatial
This is, these studies are devoted to taking the above equity measure equity, we proposed a Palma ratio to examine whether public
indicators to examine whether public transit accessibility is equitable for transit accessibility is equitable distributed in the study area.
a population group or a region. Fourth, about citizens’ trip purposes, Regarding individual spatial equity, we developed a neighboring
accessibility gap index to check which spatial units are more
advantaged or disadvantaged in the study area.
Table 1 (3) We conduct an empirical case in Kunming city, China to
Review on public traffic accessibility-based equity measure.
comprehensively examine the utility-based accessibility distri­
References PTA measure Equ. indicator Equ. Activity bution reaching business centers, transport hubs, and 3-A hos­
type
pitals, collective and individual spatial equity levels in public
Bills et al. (2022) Travel time ACS C&I Job transit accessibility, and the association between individual
Tao et al. (2022) Travel time SDC C Job spatial equity levels and some key factors including transfer times
Martens et al. (2022) Travel time ASI C&I Job
and walking distance during using public transit and the number
Jin et al. (2022) Travel time Theil index C Healthcare
Singer et al. (2022) Distance ASI I Job of neighboring traffic analysis zones.
Costa et al. (2021) Travel time TTG/TTR C&I CBD
Zuo et al. (2020) Travel time GE, Atkinson C Job The remaining of this paper is organized as follows. Section 2 pro­
indexes
vides a detailed introduction to the proposed model, including gener­
Jomehpour Chahar Travel time / C Job,
Aman & and distance Healthcare
alized travel cost function, utility-based accessibility measure,
Smith-Colin individual spatial equity measure, and collective spatial equity level.
(2020) Section 3 conducts an empirical case in Kunming city to examine the
Li and Fan (2020) Distance TGI C / spatial equity level of public transit accessibility and its associated fac­
Ermagun and Travel time / C Job, Park;
tors. Finally, Section 4 concludes this study, offers managerial insights,
Tilahun (2020)
Allen and Farber Cumulative Getis-Ord C Job and sheds light on future research directions.
(2020) opportunity statistic
Chen and Wang Cumulative- / C Job, Dining 2. Methodology
(2020) opportunity
Boisjoly et al. (2020) Travel time / C Job
Ben-Elia and Travel time Gini index C Job 2.1. Problem description and notations
Benenson (2019)
Chen et al. (2019) Cumulative- Lorenz curves C / In this section, we first formulate the generalized travel cost
opportunity and Gini formulation by public transit for each traffic analysis zone. Then based-
index
Lee and Miller Space-time Elliptic shape C Healthcare
utility accessibility measure is devised. Finally, we proposed a
(2019) prism straightforward framework for spatially collective equity measures and
Slovic et al. (2019) Travel time LISA C Job individual equity measures. Furthermore, to compare the public transit
Note: PTA – public transit accessibility; Equ. – equity; C – Collective; I – indi­ accessibility disparity among different locations within the study area,
vidual; ACS - Average consumer surplus; SDC – standard distance circle; ASI – we distinguish the old blocks and new blocks from all traffic analysis
accessibility sufficiency index; TTG &TTR – travel time gap and ratio; GE – zones. The workflow for the two-dimensional spatial equity measure of
general entropy; TGI – transit gap index; LISA – local indicators of spatial public transit accessibility is shown in Fig. 1.
association. The related notations used in the main formulations are listed below.

2
W. Li et al. Research in Transportation Economics 98 (2023) 101263

Table 2
Difference between previous studies related to accessibility measures and this study.
References PTA model TTE Price Walking Waiting Transfer Congestion

Bills et al. (2022) Utility-based SBN yes no yes yes no

Tao et al. (2022) Population-weighted SBN no yes no yes no
Martens et al. (2022) Utility-based SBN no yes no yes no
Jin et al. (2022) G2SFCA API no yes yes yes yes
Costa et al. (2021) Income-weighted API no yes yes yes yes
Zuo et al. (2020) Isochronic index SBN no yes yes no no
Ermagun and Tilahun (2020) Utility-based SBN no yes yes yes no
Boisjoly et al. (2020) Utility-based API no yes yes yes yes
Ben-Elia and Benenson (2019) Gravity-based SBN no yes yes yes no
Slovic et al. (2019) Utility-based SBN no no no no no
This study Utility-based API yes yes yes yes yes

Note: PTA – public transport accessibility; TTE – travel time estimation; API – application of program interface; G2FSCA – gaussian two-step floating catchment area.

OD pairs generation Utility estimation

Accessibility-based spatial equity Utility-based accessibility

Fig. 1. Workflow of public transit accessibility-based equity measure.

The functions of variables are explained when using them. (continued )

Sets θ Perception dispersion parameters used in the utility measure
M Set of destination type, M = {BC, TH, HP} Decision Variables
R Set of origin zones πtravel
rs,m
Travel time by transit between OD pair rs, s belonging to the mth type
Sm Set of alternative destinations of the mth type πwalk Walking time by transit between OD pair rs, s belonging to the mth type
rs,m
Nr Set of neighbors of origin zone r
πwaiting Waiting time by transit between OD pair rs, s belonging to the mth type
Tr10% Set of origin zones with top 10% accessibility rs,m

Br40% Set of destination zones with bottom 40% accessibility λticket

rs,m
Fare by transit between OD pair rs, s belonging to the mth type
Indices jm
top 10% Number of origin zones with top 10% accessibility to the mth-type
m Index of destination type destinations
r, s Index of origin, and alternative destination jm
top 10% Number of origin zones with bottom 40% accessibility to the mth-type
n Index of neighboring TAZ destinations
Parameters kNr Number of neighbors of origin zone r
μin Value of in-vehicle time (RMB/h) Functions of decision variables or parameters
μout Value of out-vehicle time (RMB/h) cmrs Travel cost by transit between OD pair rs, s belonging to the mth type
(continued on next column) (continued on next page)

3
W. Li et al. Research in Transportation Economics 98 (2023) 101263

(continued ) 2.4. Public transit accessibility-based equity

Am
r Accessibility of origin zone r to the mth-type destinations
Ar
m,neighbor Accessibility of neighbors of origin zone r to the mth-type destinations Following horizontal equity principles, this study defines public
m
Atop 10%
Average accessibility of the top 10% of the origins to the mth-type transit accessibility-based equity as the accessibility levels between a
destinations given TAZ and its neighboring TAZs should be spatially equal. We design
m
Abottom Average accessibility of the bottom 40% of the origins to the mth-type
40%
destinations
a two-dimensional measure framework for spatial equity in public
NAGm
r Neighboring accessibility gap of origin zone r to the mth-type transit accessibility, which consists of collective equity and individual
destinations equity. Collective spatial equity in public transit accessibility mainly
Pm
ratio Palma ratio of public transit accessibility related to the mth-type focuses on the accessibility gap between the richest spatial units and the
destinations
poorest spatial units, which shows whether the public transit system
provides equitable services across a geographic area. In converse, indi­
2.2. Generalized travel cost vidual spatial equity in public transit accessibility pays attention to the
accessibility gap between itself and its neighbors for each spatial unit
We assume all passengers will choose a path with the lowerest travel belonging to a geographic area, which can help find which spatial units
cost from a given TAZ to a potential destination by public transit are inequity in public transit accessibility, and their accessibility should
(El-Geneidy et al., 2016). Their generalized travel cost by public transit be improved in future public transit planning and operation.
can be formulated as Eq. (1). Eq. (1) means a door-to-door public transit
trip. The first term on the right is walking cost, πwalk
rs,m ⋅μ
out
, where we 2.4.1. Individual spatial equity
consider the walking time from the origin to the access of public transit, For individual spatial equity, measuring the disparity of public
from the egress of public transit to the destination, and the transfer time transit accessibility between a given traffic analysis zone and its
between bus and metro. The second term on the right represents the neighbors is the most straightforward approach. This approach will give
waiting cost of passengers for the first vehicle arriving at a bus stop or a specific value in spatial equity for each traffic analysis zone. It can
waiting make traffic analysis zones with inequitable accessibility easy to be
metro station (πrs,m ⋅μout ). The third term means the in-vehicle travel
found. Following the thought to measure the modal accessibility gap
cost. The final term, λticket
rs,m , is the ticket price by public transit. (Kwok & Yeh, 2004; Yang et al., 2017), this study proposes a similar
cmrs = πwalk out
+ πwaiting out
+ πtravel in ticket m
(1) neighboring accessibility gap (NAGm r ) formula to measure individual
rs,m ⋅μ rs,m ⋅μ rs,m ⋅μ + λrs,m , ∀r ∈ R, s ∈ S , m ∈ M
equity for a specific traffic analysis zone, expressed as the following
equation:
2.3. Utility-based accessibility
Amr − Am,neighbor
NAGmr = r
, ∀r ∈ R, m ∈ M (3)
Utility-based accessibility measure derives from the random utility Amr + Am,neighbor
r
theory and destination-mode choice model (Ben-Akiva & Lerman, 1979;
Domencich & McFadden, 1975). Passengers living in a given traffic 1 ∑ m
Am,neighbor = A , ∀r ∈ R, m ∈ M (4)
analysis zone, usually, will be attracted by a few alternative destinations r
krN r∈N r r
to achieve their activities in a city. It is fashionable to use systematic
utility to measure the attractiveness of these alternative destinations, It should be indicated that there are two indicators of public transit
which can consist of attributes related to alternatives, characteristics accessibility within each traffic analysis zone, including its own acces­
related to travelers, and other unobservable portions. As far as the sys­ sibility (Am
r ) and neighboring accessibility (Ar
m,neighbor
). The neighboring
tematic utility of public transit is concerned, door-to-door travel time accessibility is expressed in the average minimum expected generalized
and ticket price from the origin to destination are the most important cost from the neighboring traffic analysis zones to the same destination.
determines impacted its usage. Therefore, we take a generalized travel We define two traffic analysis zones are neighbors if their boundaries
cost function that incorporates door-to-door travel time and ticket price have an edge in common at least according to the adjacency principle, as
based on travelers’ value of time as the systematic utility of public shown in Fig. 2 (a). Eq. (3) measures the accessibility gap between the
transit. Under this context, for each TAZ, its accessibility can be regar­ given traffic analysis zone and its neighbors, and its value has been
ded as the expected minimum generalized travel cost from itself to a set standardized from − 1 to 1. If the NAGm r is between − 1 and 0, the ex­
of alternative destinations. Following a log-sum formulation (Chorus, pected minimum generalized travel cost from the given traffic analysis
2012; Handy & Niemeier, 2006), the public transit accessibility of a TAZ zone to the given destination is less than from its neighboring traffic
reaching a certain type of destination can be expressed as: analysis zones to the same destination. If the NAGm r ranges from 0 to 1,
⃒ ⃒ the expected minimum generalized travel cost from the given traffic
⃒ { }⃒ 1 ∑ ( )
Amr = E⃒⃒minm cmrs ⃒⃒ = − ln exp − θ ⋅ cmrs , ∀r ∈ R, m ∈ M (2) analysis zone to the given destination is more than that from its neigh­
θ s∈Sm boring traffic analysis zones. If the NAGm r is equal to 0, the expected
s∈S

minimum generalized travel cost has no difference whether departing

where θ is a scale parameter that reflects the perceptional variation in from the given traffic analysis zone or departing from its neighbors.
the generalized travel cost among alternative destinations. Compared to Fig. 2 (b) depicts the changes in the neighboring accessibility gap with
other measures for public transit like isochrone and gravity-based generalized travel costs. In summary, the more the neighboring acces­
approach, an advantage of the utility-based approach is that it can sibility gap approaches zero, the more equitable the services provided by
capture the impact of the travel mode on accessibility and equity by the public transit system are. Without loss of generality, we define the
using the generalized travel cost function. Another advantage is that the following mapping relationships between the neighboring accessibility
introduction of the expected minimum generalized travel cost makes the gap and inequity level, as shown in Table 3.
results of the accessibility measure easier to interpret in intuition. Due to
that travelers usually expect as less travel cost as possible, therefore, the 2.4.2. Collective spatial equity
higher value of utility-based accessibility means poorer public transit Inspired by Palma (2011) who applied an index to investigate in­
accessibility. come inequity between the income share of the top group and the bot­
tom group, this study also adopts a similar ratio index to formulate
collective equity in public transit accessibility based on Eq. (5):

4
W. Li et al. Research in Transportation Economics 98 (2023) 101263

(a) (b)

Neighboring accessibility gap

Generalized travel cost

Fig. 2. Neighboring traffic analysis zones and neighboring accessibility gap.

Kunming had respectively reached 1.37 and 0.61 million passengers per
Table 3
day. Fig. 3 (a) plots the location of Kunming and the spatial distribution
Mapping relationships between the neighboring accessibility gap and inequity
of public transit.
level.
We use the traffic analysis zone as a spatial unit to measure public
NAG [ − 0.05, [ − 0.25, − 0.05) ∪ [ − 0.5, − 0.25) ∪ [ − 1, − 0.5) ∪ transit accessibility and its equity. Based on the urban road network of
0.05] (0.05, 0.25] (0.25, 0.5] (0.5, 1]
Kunming, the study area has been divided into 524 traffic analysis zone.
Level no slight inequity medium inequity severe For each zone, we assume that all public transit users will depart from its
inequity inequity
centroid, then ride the transit system to reach their destinations. Fig. 3
Note: NAG – neighboring accessibility gap. (b) shows the spatial distribution of traffic analysis zones used in this
study.
∑
m
1
Amr Regarding possible destinations related to transit-based trips, we
jm
Atop top 10% r
r∈T10% only consider cities’ three daily activities, including leisure activities,
(5)
10%
Pmratio = = ∑ , ∀m ∈ M
seeking medical care, and transfer of intercity travel. Correspondingly,
m 1
Abottom 40% jm
Amr
we select three types of POIs (points of interest) as destinations: the
bottom 40% r∈Br40%

In this Paloma ratio model of public transit accessibility, the business center POI (comprehensive commercial buildings where citi­
numerator denotes the top 10%‘s average accessibility over the whole zens can shop, repast, visit, and recreate), the medical POI (3-A hospitals
area, and the denominator is the lowest 40%‘s average accessibility where citizens can seek medical care), the transport hub POI (airport or
across the area. Unlike the frequently used Gini index that includes the railway station where citizens can transfer for intercity travel). There are
distribution over all groups, for example, all populations in a country, or 21 business centers, 19 3-A hospitals, and 3 transport hubs within the
all spatial units (traffic analysis zones) in a city, the Paloma ratio index is study area by searching for the Gaode map (also named the Amap1), as
more concerned about extreme groups (the richest and the poorest) shown in Fig. 3 (c)–(e).
excluding the groups with middle accessibility. In the context of spatial
accessibility, using the Paloma ratio rather than other measures is 3.2. OD pairs generation
because of the ease of interpretability and communication, which
directly shows how much the richest accessibility is higher than the Based on the identified traffic analysis zones and possible destina­
poorest. This just fills a drawback that the Gini index is not very sensitive tions, the origin and destination (OD) pairs between them are extracted
to measure differences in the bottom portion of the distribution. to estimate the OD travel time and trip fare for public transit
Therefore, the Paloma ratio has been gradually adopted in existing accessibility-based equity research in this paper. It should be indicated
studies related to assessing accessibility-based equity (Guzman & we only consider the most possible public transit trips rather than all
Oviedo, 2018; Liu et al., 2021; Rosas-Satizábal et al., 2020). trips. For example, for leisure activities, we only examine the public
transit trips from each traffic analysis zone to business centers with the
3. Case study top 5 shortest travel times. The process of seeking medical care is the
same. Due to only three transport hubs in the study area, for public
3.1. Study area transit-based intercity travel, we check all public transit trips from each
traffic analysis zone to each transport hub. Table 4 describes the number
The main urban area of Kunming city is used as an empirical area in of OD pairs used for public transit accessibility-based equity modeling.
this study. As the capital of Yunnan Province located in southwestern Then, the Application Programming Interface (API) of the Gaode
China, Kunming city has a population of 3.58 million inhabitants and map LBS (Location Based Service) open platform2 has been adopted to
covered an area of 3, 243 km2. It owns a convenient, dense, and efficient automatically query trip information including travel distance,
public transit system that serves citizens’ daily trips by 280 lines and
10,136 bus stops. Thus, it is also one of the outstanding transit metro­
polises in China. By 2020, the bus ridership and metro ridership in 1
https://www.amap.com/.
2
https://lbs.amap.com/.

5
W. Li et al. Research in Transportation Economics 98 (2023) 101263

Fig. 3. Study area: Kunming city, China.

Table 4 Table 5
OD pairs information. Three samples of trip information of all OD pairs.
Trip scenarios Leisure Seeking medical Transfer of intercity OD id TAZ_154 – TAZ_132 – TAZ_175 –
activities care travel BC_21 3AH_18 TH_3

Origin points 524 TAZs 524 TAZs 524 TAZs Price (RMB) 1.0 4.0 7.0
Destinations 28 business 19 3-A hospitals 3 transport hubs Distance (km) 5.72 17.29 42.68
points centers Duration (minutes) 38.97 78.75 91.53
Effective Top 5 Top 5 Top 3 In-bus time (minutes) 28.77 50.76 64.85
destinations Out-bus time 10.20 27.97 26.64
All OD pairs 14,672 9956 1572 (minutes)
Effective OD pairs 2620 2620 1572 Walking distance 0.71 1.96 1.87
(km)
Note: O – origin; D – destination; TAZ – traffic analysis zone. Transfer times 0 1 2

Note: OD – origin and destination; TAZ – traffic analysis zone; BC – business

duration, and fare for each traffic analysis zone in Python script. The API
center; 3AH – 3-A hospital; TH – transport hub.
inputs location coordinates (latitude and longitude) of each traffic
analysis zone and planned departure time, and outputs corresponding
3.3. Unity-based accessibility distribution
trip data based on the road network, transit operation schedule, and
real-time traffic of the Gaode map. In our study, the location information
Before discussing public transit accessibility-based equity, we first
of traffic analysis zones has been prepared in advance based on ArcGIS
need to preset a few exogenous parameters to assist examine different
software, and the planned departure time is set at 10 a.m. on August 1st.
travel scenarios, including the perceptional variation parameter (θ), the
Compared with other methods, for example, using the Network Analysis
tool in ArcGIS or distance and speed-based measures, the Gaode map’s value of in-vehicle time (μin ), and the value of out-vehicle time (μout ).
API has three obvious advantages: (1) the road network used for travel First, the perceptional variation parameter is set as 0.5. Referring to Liu
time estimation is more complete than the self-built road network. (2) et al. (2022), the following formula is used to determine the value of
the Gaode map’s API introduced the transit schedule which might make in-vehicle time:
/ / / /
the estimation of waiting time more precise. (3) The real-time speed and μin = W 12 4.2 5 8 (6)
congestion have also been considered. Sample trip information is shown
in Table 5. where W denotes the Gross Domestic Product (GDP) per capita in the
study area, 12, 4.2, 5, and 8 respectively are the number of months per

6
W. Li et al. Research in Transportation Economics 98 (2023) 101263

year, the number of weeks per month, the number of working days per which makes convenient public transit accessibility to transport hubs
week, and the number of working hours per day. In our study, the value available in these areas, providing affordable and also efficient services.
of time is assumed to be homogeneous for all users. The effect of het­
erogeneous users is left for future study. According to the GDP per capita 3.4. Collective spatial equity
level in Kunming city in 2021, the value of in-vehicle time is set as 40
RMB per, and the out-vehicle VOT is set as 1.5 times the value of in- To calculate the public transit accessibility-based Palma ratio, we
vehicle time. classified each traffic analysis zone by accessibility decile considering
Fig. 4 presents the spatial distribution of public transit accessibility the expected minimum generalized travel cost. Each decile is composed
in all traffic analysis zones. Considering the expected minimum gener­ of a different number of traffic analysis zones. The top 10% of public
alized travel cost in routes, citizens in a part of traffic analysis zones that transit accessibility is concentrated in 53 traffic analysis zones, and the
are either close to the business centers, transport hubs, or 3-A hospitals poorest 40% of public transit accessibility is distributed among 210
or directly linked by transit lines (inappreciable walking) can arrive at traffic analysis zones across the whole study area. Fig. 6 presents the
them within 30-RMB travel cost by public transit. Citizens living in the statistical distribution of the accessibility-based Palma ratio in Kunming
majority of traffic analysis zones can reach business centers, transport city when taking into account distinct destinations and locations.
hubs, and 3-A hospitals within 30-RMB to 90-RMB generalized travel Fig. 6 (a) shows how the magnitude of public transit accessibility
costs. For some traffic analysis zones, however, citizens’ generalized inequity among traffic analysis zones with rich accessibility and poor
travel costs are more than 90 RMB away from them. Most of these traffic accessibility in Kunming city when considering different destinations.
analysis zones are the furthest from business centers, transport hubs, and The public transit accessibility-based Palma ratios for business centers,
3-A hospitals or are located in suburban and semi-rural areas where transport hubs, and 3-A hospitals are respectively 0.40, 0.41, and 0.38.
there is inadequate coverage of public transit services. They look very close, the value of which means that the expected min­
Fig. 5 (a) shows a significant disparity in public transit accessibility imum generalized travel cost from the poorest 40% of traffic analysis
of all traffic analysis zones. As far as all traffic analysis zones are con­ zones to these destinations is almost 2.5 times (1/0.4) as much as from
cerned, citizens to transport hubs usually bear the highest travel cost. the richest 10% of traffic analysis zones. This should be noticed by
Specifically, the mean public transit accessibility to business centers, public transit planners and managers because the disadvantages in
transport hubs, and 3-A hospitals are respectively 30 RMB, 45 RMB, and utility-based accessibility for the poorest 40% can lead to deterioration
34 RMB. The expected minimum generalized travel cost to transport in socioeconomic status. It is also worth noting that there are no sig­
hubs is more than 50% to business centers, and 32% to 3-A hospitals. nificant public transit accessibility gaps among reaching business cen­
The public transit accessibility to business centers and 3-A hospitals are ters, transport hubs, and 3-A hospitals when considering the collective
very close. A possible reason is that the number of transport hubs in Palma ratio.
Kunming city is relatively less than the number of business centers and Moreover, the disparity of collective accessibility-based equity
3-A hospitals. In our study, we only considered three key transport hubs among different locations in the study area is compared, as shown in
in the study area, e.g., the Kunming Changshui International Airport, Fig. 6 (b)–(d). As expected, collective spatial equity levels of public
Kunming Railway Station, and Kunmingnan Railway Station. transit accessibility for reaching each kind of destination seem to be
As shown in Fig. 5 (b), the statistical distribution of public transit changed among different urban locations. A consistent finding can be
accessibility for three types of locations, including old blocks, new seen for three types of destinations, public transit accessibility-based
blocks, and other locations, is significantly different. Recalling that the Palma ratio in old blocks are both higher than in new blocks. For
old blocks are located in the city center, and the new blocks are located example, for reaching business centers by public transit, the
in the south of Kunming city. In terms of public transit accessibility to accessibility-based Palma ratio in old blocks and new blocks are
business centers and 3-A hospital, their accessibility distributions are respectively 0.62 and 0.46. This result indicates that for citizens in old
consistent that both follow such a pattern: old blocks > new blocks > blocks, it will take 1.61 times (1/0.62) the expected minimum gener­
other locations. Interestingly, the public transit accessibility to transport alized travel cost departing from the poorest 40% of traffic analysis
hubs is the opposite. It can be found the accessibility to transport hubs zones to reach business centers than departing from the richest 10% of
from new blocks is superior to old blocks and other locations. This is ones. However, for citizens in old blocks, the gap in expected minimum
attributed to the Kunmingnan railway station located in new blocks, generalized travel cost between the top 10% of traffic analysis zones and

Fig. 4. Spatial distribution of public transit accessibility in TAZs.

7
W. Li et al. Research in Transportation Economics 98 (2023) 101263

Fig. 5. Disparity of public transit accessibility among different locations.

Fig. 6. Statistic distribution of the public transit accessibility-based Palma ratio.

the bottom 40% of traffic analysis zones can reach 2.17 (1/0.46) times. 3.5. Individual spatial equity
A similar situation exists in the public transit accessibility-based Palma
ratio reaching transport hubs and 3-A hospitals. That is to say, traffic The collective Palma ratio in section 4.3 uncovers that public transit
analysis zones in new blocks may face inequity in public transit acces­ accessibility may be inequitable distributed in the study area. However,
sibility relative to old blocks. A possible explanation is that there is it does not indicate which traffic analysis zones maybe face poorer
complete coverage in public transit networks and stops in the old blocks. public transit accessibility. Therefore, following the neighboring acces­
However, in new blocks, some feeding transit networks need further sibility gap index model (see Eqs. (3) and (4)), this section shows the
improvement. individual equitable level for each traffic analysis zone. We first drew
the scatter diagram and histogram of its public transit accessibility and
its neighboring public transit accessibility for each traffic analysis zone

8
W. Li et al. Research in Transportation Economics 98 (2023) 101263

(see Fig. 7). Fig. 7 shows that the traffic analysis zones’ public transit 3.6. Analysis of associated factors
accessibility is not far greater than that of their neighbors. The vast
majority of scattered points are distributed on both sides of the gray After showing the spatial pattern of individual spatial equity in
dotted line (y = x), meaning that the expected minimum generalized public transit accessibility, we further examine whether the transfer
travel cost by public transport for each traffic analysis zone is likely to its times and walking distance during riding public transit and the number
neighboring zones. This is, the public transit accessibility is equitable for of neighboring traffic analysis zones are associated with the individual
these traffic analysis zones. However, a few scattered points are falling spatial equity level. Fig. 9 depicts the scatter diagram and regression plot
on the lower part of the gray dotted line (y = x), indicating that the gap of the above factors and the neighboring accessibility gap for each traffic
in citizens’ expected minimum generalized travel cost between given analysis zone. Correspondingly, it can be first observed that the neigh­
traffic analysis zones and their neighbors becomes large. boring accessibility gap seems to be not directly associated with transfer
Furthermore, based on the mapping relationships between the times during the whole trip as the regression plots almost overlap with
neighboring accessibility gap and the inequity level of public transit the red dotted baseline (y = 0). Interestingly, it can be found that the
accessibility (see Table 1), the histogram distribution and spatial dis­ neighboring accessibility gap can gradually increase with the growth of
tribution of the neighboring accessibility gap values for each traffic walking distance during taking public transit. When the walking dis­
analysis zone are shown in the following (see Fig. 8). The value of the tance is within 2 km, the neighboring accessibility gap is very small.
neighboring accessibility gap ranges from − 0.4 to 0.4 under scenarios That is to say, the traffic analysis zones are spatially equitable in public
reaching business centers, transport hubs, and 3-A hospitals. Overall, transit accessibility under this scenario. However, a long walking dis­
public transit accessibility for given traffic zones and their neighbors tance out of travelers’ acceptable walkability can lead to some spatial
does not present a huge gap because of around 300 traffic analysis zones inequities in public transit accessibility. The number of neighboring
(almost 57%) with a neighboring accessibility gap index between − 0.05 traffic analysis zones also presents a similar association with individual
and 0.05. These traffic analysis zones are almost located in new blocks spatial equity in public transit accessibility. But the growth affected by
and old blocks of Kunming city. The number of traffic analysis zones the number of neighboring traffic analysis zones is obviously less than
with values of neighboring accessibility gaps within an interval of − 0.25 the ones affected by walking distance. The traffic analysis zones with
and − 0.05 and an interval of 0.05 and 0.25 was relatively medium, 5–8 neighbors usually have more equitable public transit accessibility.
accounting for around 38%. These traffic analysis zones with slight
inequity are mainly concentrated in the peripheral urban areas and 4. Conclusions
urban-rural fringe areas, which indicates that the layout of transit net­
works and stops in these areas should be paid attention to improve Due to lacking related studies in both examining individual spatial
spatial inequity in public transit accessibility. In addition, a few traffic equity and collective spatial equity in public transit accessibility, we
analysis zones are with medium inequity in public transit accessibility, hope to obtain some referred insights for developing helpful policies on
the number of which only occupied less than 5% of all traffic analysis equitable public transit planning and operation. This research develops
zones in the study area. These zones are distributed in the urban fringe a novel measure framework for public transit accessibility that in­
area where the public transit networks have little service catchment. tegrates individual equity and collective equity to cope with such two
meaningful questions: how is the spatial equity in public transit acces­
sibility for the whole region and which zones’ public transit accessibility

Fig. 7. The scatter diagram and histogram of neighboring accessibility gaps.

9
W. Li et al. Research in Transportation Economics 98 (2023) 101263

Fig. 8. The spatial distribution of inequity levels.

is more disadvantaged in the region? generalized travel cost by public transit in these zones is nearly
Considering a multimodal transport network covering walking, bus, equal to the ones in their neighboring traffic analysis zones. The
and metro, a utility-based accessibility measure model embedded with number of traffic analysis zones with slight inequity in public
the log-sum formula is proposed to quantitatively evaluate spatial transit accessibility accounts for around 38%, which are mainly
accessibility from each traffic analysis zone to possible activity desti­ located in the peripheral urban areas and urban-rural fringe
nations. The utility function of accessibility measure encompasses the areas. Only less than 5% of traffic analysis zones are with medium
generalized travel cost by public transit considering travelers’ travel inequity in public transit accessibility.
time, waiting time, transfer time, and ticket price. They are connected by (4) Fourth, we also found that the individual spatial inequity level in
travelers’ value of time. We then apply a Palma ratio of public transit public transit accessibility can be associated with the walking
accessibility to examine the accessibility gap between the top 10% of distance during taking public transit and the number of neigh­
traffic analysis zones and the bottom 40% of traffic analysis zones, boring traffic analysis zones. However, a direct connection be­
which is defined as the collective spatial equity in our study. Closely tween the individual spatial inequity level and the transfer times
followed by this, we design a neighboring accessibility gap index to during using transit is not been obviously observed.
measure individual equity levels in public transit accessibility. Finally,
an empirical study of Kunming city is conducted to analyze the present For policymakers working to improve spatial inequity in public
equity situation in public transit accessibility individually and collec­ transit networks and increase public transit users, this research provides
tively, including accessibility reaching business centers, transport hubs, some referable insights. While encouraging public transit use, seamless
and 3-A hospitals. Additionally, the relationships between individual connection for bus and metro and excellent walkability between origin/
spatial equity and transfer times, walking distance, and the number of destination points and nearest bus stops may be efficient measures to
traffic analysis zones are also checked. By considering our results, increase public transit ridership, as they may be associated with the
several conclusions can be drawn as follows. expected minimum travel cost by transit. The high levels of neighboring
accessibility gaps in some traffic analysis zones indicate that many cit­
(1) First, in the aspect of utility-based accessibility measures, there is izens lack convenient alternatives to public transit use. Our results
a significant disparity in public transit accessibility of all traffic suggested that this risk can be somewhat easily found in urban periph­
analysis zones. The expected minimum generalized travel cost of eries and transport poverty areas. These findings can be useful for pol­
reaching transport hubs by public transit is more than 50% to icymakers to design equitable transit networks and attract more users to
business centers, and 32% to 3-A hospitals in Kunming city. The shift the car mode to transit mode. On the one hand, authorities and
public transit accessibility reaching business centers and 3-A agencies can take ride-hailing or taxi services as a feeder to connect with
hospitals in old blocks are superior to the new blocks. the transit network (Singer et al., 2022; Yan et al., 2019; Zhang & Khani,
Conversely, the accessibility to transport hubs is more advan­ 2021; Zuniga-Garcia et al., 2022). Incorporating ride-hailing and transit
taged in new blocks. services is conducive to decreasing passengers’ waiting time costs in
(2) Second, in terms of collective spatial equity in public transit these disadvantaged traffic analysis zones. On the other hand, this result
accessibility, accessibility-based Palma ratios reaching business can motivate operators and managers to optimize transit networks with
centers, transport hubs, and 3-A hospitals are both very close to the schedule- and frequency-based services in accessibility-equitable
0.4, which means that in Kunming city, the expected minimum and accessibility-inequitable traffic analysis zones respectively (Jiang
generalized travel cost by public transit from the poorest 40% of et al., 2022).
traffic analysis zones to these destinations is almost 2.5 times as Several extensions under this study need to be carried out in the
much as from the richest 10% of traffic analysis zones. For three future. For example, (1) this study mainly involved three daily activities:
types of destinations, the public transit accessibility-based Palma leisure activities, seeking medical care, and intercity travel correlated
ratio suggested that the collective spatial equity in old blocks is with public transit. The public transit accessibility-based equity of other
superior to new blocks. activities, such as commute activities and city tourism activities also is
(3) Third, concerning individual spatial equity in public transit worth noticing and comparing. (2) In utility-based accessibility mea­
accessibility, almost 57% of traffic analysis zones have no sures, the value of time (VOT) is assumed to be homogeneous for all
obvious spatial inequity because the expected minimum traffic analysis zones. One way to improve this would be to incorporate

10
W. Li et al. Research in Transportation Economics 98 (2023) 101263

Fig. 9. The scatter diagram and regression plot of some factors and neighboring accessibility gaps.

spatially heterogeneous VOT into the accessibility measure. (3) In accessibility for designing the equitable public transit system.
spatial equity measure, we do not consider the effects of zonal popula­
tion and economics. The following studies can involve these factors to CRediT authorship contribution statement
develop a population/economics-weighted Palma ratio and neighboring
accessibility gap index in spatial equity measure. (4) With a zonal-level Wu Li: Conceptualization, Methodology, Software, Investigation,
unit of analysis, the results of the accessibility measure and spatial eq­ Validation, Visualization, Writing - original draft, Writing - review &
uity measure in traffic analysis zones may be inevitably sensitive to the editing. Haotian Guan: Validation, Writing - review & editing. Wen­
modifiable areal unit problem (MAUP). Although the generation of wen Qin: Methodology, Validation, Writing - review & editing. Xiao­
traffic analysis zones in this study is consistent with the layout of the feng Ji: Supervision, Funding acquisition, Conceptualization, Project
road network in Kunming city, other spatial units like Voronoi diagram administration, Writing - review & editing.
cells and hexagonal cells are worth to been introduced in future studies
to relieve MAUP bias. (5) Another interesting topic is to examine how
the built environment affects individual spatial equity in public transit

11
W. Li et al. Research in Transportation Economics 98 (2023) 101263

Declaration of competing interest Guzman, L. A., & Oviedo, D. (2018). Accessibility, affordability and equity: Assessing
‘pro-poor’ public transport subsidies in Bogotá. Transport Policy, 68, 37–51. https://
doi.org/10.1016/j.tranpol.2018.04.012
No conflict of interest exists in the submission of this manuscript, and Handy, S. L., & Niemeier, D. A. (2006). Measuring accessibility: An exploration of issues
that is approved by all authors for publication. I would like to declare on and alternatives. Environment and Planning A: Economy and Space, 29(7), 1175–1194.
behalf of my co-authors that the work described was original research https://doi.org/10.1068/a291175
Hansen, W. G. (2007). How accessibility shapes land use. Journal of the American Institute
that has not been published previously, and is not under consideration of Planners, 25(2), 73–76. https://doi.org/10.1080/01944365908978307
for publication elsewhere, in whole or in part. All the authors listed have Ignaccolo, M., Inturri, G., Giuffrida, N., & Torrisi, V. (2016). Public transport accessibility
approved the manuscript that is enclosed. and social exclusion: Making the connections. In Proceedings of the third international
conference on traffic and transport engineering (ICTTE) (pp. 24–25). Belgrade, Serbia.
Jiang, Y., Rasmussen, T. K., & Nielsen, O. A. (2022). Integrated optimization of transit
Data availability networks with schedule- and frequency-based services subject to the bounded
stochastic user equilibrium. Transportation Science. https://doi.org/10.1287/
trsc.2022.1148
Data will be made available on request. 28 Jin, T., Cheng, L., Wang, K., Cao, J., Huang, H., & Witlox, F. (2022). Examining equity
in accessibility to multi-tier healthcare services across different income households
Acknowledgement using estimated travel time. Transport Policy, 121, 1–13. https://doi.org/10.1016/j.
tranpol.2022.03.014.
Jomehpour Chahar Aman, J., & Smith-Colin, J. (2020). 3.Transit Deserts: Equity analysis
This study is funded by the National Natural Science Foundation of of public transit accessibility. Journal of Transport Geography, 89, Article 102869.
China (No.41501174 and No.71904068) and Innovation Guidance and https://doi.org/10.1016/j.jtrangeo.2020.102869
Kelobonye, K., McCarney, G., Xia, J., Swapan, M. S. H., Mao, F., & Zhou, H. (2019).
Enterprise Cultivation Planning Project of Yunnan Province of China Relative accessibility analysis for key land uses: A spatial equity perspective. Journal
(No.202004AR040022). We would like to appreciate the Gaode Map of Transport Geography, 75, 82–93. https://doi.org/10.1016/j.jtrangeo.2019.01.015
(AMAP) for providing us with free travel time estimation by public Kwok, R. C. W., & Yeh, A. G. O. (2004). The use of modal accessibility gap as an indicator
for sustainable transport development. Environment and Planning A: Economy and
transit.
Space, 36(5), 921–936. https://doi.org/10.1068/a3673
Lee, J., & Miller, H. J. (2019). 16.Analyzing collective accessibility using average space-
References time prisms. Transportation Research Part D: Transport and Environment, 69, 250–264.
https://doi.org/10.1016/j.trd.2019.02.004
Lee, R. J., Sener, I. N., & Jones, S. N. (2016). Understanding the role of equity in active
Allen, J., & Farber, S. (2020). 12.Planning transport for social inclusion: An accessibility-
transportation planning in the United States. Transport Reviews, 37(2), 211–226.
activity participation approach. Transportation Research Part D: Transport and
https://doi.org/10.1080/01441647.2016.1239660
Environment, 78, Article 102212. https://doi.org/10.1016/j.trd.2019.102212
Li, Y., & Fan, W.“D. (2020). 4.Modeling and evaluating public transit equity and
Behbahani, H., Nazari, S., Jafari Kang, M., & Litman, T. (2018). A conceptual framework
accessibility by integrating general transit feed specification data: Case study of the
to formulate transportation network design problem considering social equity
city of charlotte. Journal of Transportation Engineering, Part A: Systems, 146(10),
criteria. Transportation Research Part A: Policy and Practice, 125, 171–183. https://
Article 04020112. https://doi.org/10.1061/JTEPBS.0000426
doi.org/10.1016/j.tra.2018.04.005
Li, L., Ren, H., Zhao, S., Duan, Z., Zhang, Y., & Zhang, A. (2017). Two dimensional
Ben-Akiva, M., & Lerman, S. R. (1979). Disaggregate travel and mobility-choice models
accessibility analysis of metro stations in Xi’an, China. Transportation Research Part
and measures of accessibility. In Behavioural travel modelling. Routledge.
A: Policy and Practice, 106, 414–426. https://doi.org/10.1016/j.tra.2017.10.014
Ben-Elia, E., & Benenson, I. (2019). 2.A spatially-explicit method for analyzing the equity
Litman, T. (2012). Evaluating transportation equity: Guidance for incorporating distributional
of transit commuters’ accessibility. Transportation Research Part A: Policy and
impacts in transport planning. Victoria, Australia: Victoria Transport Policy Institute.
Practice, 120, 31–42. https://doi.org/10.1016/j.tra.2018.11.017
Liu, D., Kwan, M.-P., & Kan, Z. (2021). Assessing job-access inequity for transit-based
Bills, T. S., Twumasi-Boakye, R., Broaddus, A., & Fishelson, J. (2022). 6.Towards transit
workers across space and race with the Palma ratio. Urban Research and Practice,
equity in Detroit: An assessment of microtransit and its impact on employment
1–27. https://doi.org/10.1080/17535069.2021.1923795
accessibility. Transportation Research Part D: Transport and Environment, 109, Article
Liu, Q., Tao, Z., & Zhou, J. (2022). Assessing equitable accessibility based on total travel
103341. https://doi.org/10.1016/j.trd.2022.103341
cost and multiple options: Case study of airport ground access in Hong Kong. Journal
Boisjoly, G., Serra, B., Oliveira, G. T., & El-Geneidy, A. (2020). 24.Accessibility
of Urban Planning and Development, 148(1), Article 04021061. https://doi.org/
measurements in são paulo, rio de Janeiro, curitiba and recife, Brazil. Journal of
10.1061/(ASCE)UP.1943-5444.0000797
Transport Geography, 82, Article 102551. https://doi.org/10.1016/j.
Lucas, K., van Wee, B., & Maat, K. (2015). A method to evaluate equitable accessibility:
jtrangeo.2019.102551
Combining ethical theories and accessibility-based approaches. Transportation, 43
Camporeale, R., Caggiani, L., Fonzone, A., & Ottomanelli, M. (2016). Quantifying the
(3), 473–490. https://doi.org/10.1007/s11116-015-9585-2
impacts of horizontal and vertical equity in transit route planning. Transportation
Martens, K., Golub, A., & Robinson, G. (2012). A justice-theoretic approach to the
Planning and Technology, 40(1), 28–44. https://doi.org/10.1080/
distribution of transportation benefits: Implications for transportation planning
03081060.2016.1238569
practice in the United States. Transportation Research Part A: Policy and Practice, 46
Chen, Y., Bouferguene, A., Shen, Y., & Al-Hussein, M. (2019). 8.Assessing accessibility-
(4), 684–695. https://doi.org/10.1016/j.tra.2012.01.004
based service effectiveness (absev) and social equity for urban bus transit: A
Martens, K., Singer, M. E., & Cohen-Zada, A. L. (2022). 25.Equity in accessibility: Moving
sustainability perspective. Sustainable Cities and Society, 44, 499–510. https://doi.
from disparity to insufficiency analyses. Journal of the American Planning Association.
org/10.1016/j.scs.2018.10.003
https://doi.org/10.1080/01944363.2021.2016476, 1–16.
Chen, N., & Wang, C.-H. (2020). 19.Does green transportation promote accessibility for
Nahmias-Biran, B., Sharaby, N., & Shiftan, Y. (2013). Equity aspects in transportation
equity in medium-size U.S. cites? Transportation Research Part D: Transport and
projects: Case study of transit fare change in haifa. International Journal of Sustainable
Environment, 84, Article 102365. https://doi.org/10.1016/j.trd.2020.102365
Transportation, 8(1), 69–83. https://doi.org/10.1080/15568318.2012.758525
Chorus, C. G. (2012). Logsums for utility-maximizers and regret-minimizers, and their
Niehaus, M., Galilea, P., & Hurtubia, R. (2016). Accessibility and equity: An approach for
relation with desirability and satisfaction. Transportation Research Part A: Policy and
wider transport project assessment in Chile. Research in Transportation Economics, 59,
Practice, 46(7), 1003–1012. https://doi.org/10.1016/j.tra.2012.04.008
412–422. https://doi.org/10.1016/j.retrec.2016.05.003
Costa, C., Ha, J., & Lee, S. (2021). 36.Spatial disparity of income-weighted accessibility
Oswald Beiler, M., & Mohammed, M. (2016). Exploring transportation equity:
in Brazilian cities: Application of a google maps API. Journal of Transport Geography,
Development and application of a transportation justice framework. Transportation
90, Article 102905. https://doi.org/10.1016/j.jtrangeo.2020.102905
Research Part D: Transport and Environment, 47, 285–298. https://doi.org/10.1016/j.
Delmelle, E. C., & Casas, I. (2012). Evaluating the spatial equity of bus rapid transit-based
trd.2016.06.007
accessibility patterns in a developing country: The case of Cali, Colombia. Transport
Palma, J. G. (2011). Homogeneous middles vs. Heterogeneous tails, and the end of the
Policy, 20, 36–46. https://doi.org/10.1016/j.tranpol.2011.12.001
‘inverted-U’: It’s all about the share of the rich. Development and Change, 42(1),
Domencich, T. A., & McFadden, D. (1975). Urban travel demand; a behavioural analysis.
87–153. https://doi.org/10.1111/j.1467-7660.2011.01694.x
https://trid.trb.org/view/1175810.
Ricciardi, A. M., Xia, J., & Currie, G. (2015). Exploring public transport equity between
El-Geneidy, A., Levinson, D., Diab, E., Boisjoly, G., Verbich, D., & Loong, C. (2016). The
separate disadvantaged cohorts: A case study in perth, Australia. Journal of Transport
cost of equity: Assessing transit accessibility and social disparity using total travel
Geography, 43, 111–122. https://doi.org/10.1016/j.jtrangeo.2015.01.011
cost. Transportation Research Part A: Policy and Practice, 91, 302–316. https://doi.
Rosas-Satizábal, D., Guzman, L. A., & Oviedo, D. (2020). Cycling diversity, accessibility,
org/10.1016/j.tra.2016.07.003
and equality: An analysis of cycling commuting in Bogotá. Transportation Research
Ermagun, A., & Tilahun, N. (2020). 5.Equity of transit accessibility across Chicago.
Part D: Transport and Environment, 88, Article 102562. https://doi.org/10.1016/j.
Transportation Research Part D: Transport and Environment, 86, Article 102461.
trd.2020.102562
https://doi.org/10.1016/j.trd.2020.102461
Singer, M. E., Cohen-Zada, A. L., & Martens, K. (2022). Core versus periphery: Examining
Golub, A., & Martens, K. (2014). Using principles of justice to assess the modal equity of
the spatial patterns of insufficient accessibility in U.S. metropolitan areas. Journal of
regional transportation plans. Journal of Transport Geography, 41, 10–20. https://doi.
Transport Geography, 100, Article 103321. https://doi.org/10.1016/j.
org/10.1016/j.jtrangeo.2014.07.014
jtrangeo.2022.103321
Grengs, J. (2013). Nonwork accessibility as a social equity indicator. International Journal
Slovic, A. D., Tomasiello, D. B., Giannotti, M., Andrade, M. de F., & Nardocci, A. C.
of Sustainable Transportation, 9(1), 1–14. https://doi.org/10.1080/
(2019). 85.The long road to achieving equity: Job accessibility restrictions and
15568318.2012.719582

12
W. Li et al. Research in Transportation Economics 98 (2023) 101263

overlapping inequalities in the city of São Paulo. Journal of Transport Geography, 78, Journal of Transport Geography, 60, 21–32. https://doi.org/10.1016/j.
181–193. https://doi.org/10.1016/j.jtrangeo.2019.06.003 jtrangeo.2017.02.005
Stanley, J., & Lucas, K. (2008). Social exclusion: What can public transport offer? Yan, X., Levine, J., & Zhao, X. (2019). Integrating ridesourcing services with public
Research in Transportation Economics, 22(1), 36–40. https://doi.org/10.1016/j. transit: An evaluation of traveler responses combining revealed and stated
retrec.2008.05.009 preference data. Transportation Research Part C: Emerging Technologies, 105, 683–696.
Tahmasbi, B., & Haghshenas, H. (2019). Public transport accessibility measure based on https://doi.org/10.1016/j.trc.2018.07.029
weighted door to door travel time. Computers, Environment and Urban Systems, 76, Zhang, Y., & Khani, A. (2021). Integrating transit systems with ride-sourcing services: A
163–177. https://doi.org/10.1016/j.compenvurbsys.2019.05.002 study on the system users’ stochastic equilibrium problem. Transportation Research
Tao, S., He, S., Ettema, D., & Luo, S. (2022). 7.The role of transit accessibility in Part A: Policy and Practice, 150, 95–123. https://doi.org/10.1016/j.tra.2021.05.008
influencing the activity space and non-work activity participation of different Zhou, Q., Dai, D., Wang, Y., & Fan, J. (2018). Decade-long changes in disparity and
income groups. Journal of Transport and Land Use, 15(1). https://doi.org/10.5198/ distribution of transit opportunity in shenzhen China: A transportation equity
jtlu.2022.2075 perspective, 2018 Journal of Advanced Transportation, 1–16. https://doi.org/
Wang, J., Xu, C., Pauleit, S., Kindler, A., & Banzhaf, E. (2019). Spatial patterns of urban 10.1155/2018/7127342.
green infrastructure for equity: A novel exploration. Journal of Cleaner Production, Zuniga-Garcia, N., Gurumurthy, K. M., Yahia, C. N., Kockelman, K. M., &
238, Article 117858. https://doi.org/10.1016/j.jclepro.2019.117858 Machemehl, R. B. (2022). Integrating shared mobility services with public transit in
Welch, T. F. (2013). Equity in transport: The distribution of transit access and areas of low demand. Journal of Public Transportation, 24, Article 100032. https://
connectivity among affordable housing units. Transport Policy, 30, 283–293. https:// doi.org/10.1016/j.jpubtr.2022.100032
doi.org/10.1016/j.tranpol.2013.09.020 Zuo, T., Wei, H., Chen, N., & Zhang, C. (2020). 1. First-and-last mile solution via
Yang, W., Chen, B. Y., Cao, X., Li, T., & Li, P. (2017). The spatial characteristics and bicycling to improving transit accessibility and advancing transportation equity.
influencing factors of modal accessibility gaps: A case study for guangzhou, China. Cities, 99, Article 102614. https://doi.org/10.1016/j.cities.2020.102614

13