Towards a cycling-friendly cityAn updated review of the associations between built environment and cycling behaviors(2007–2017)Yang, Yiyang; Wu, Xueying; Zhou, Peiling; Gou, Zhonghua; Lu, Yi

Published in:Journal of Transport and Health

Published: 01/09/2019

Document Version:Final Published version, also known as Publisher’s PDF, Publisher’s Final version or Version of Record

License:CC BY-NC-ND

Publication record in CityU Scholars:Go to record

Published version (DOI):10.1016/j.jth.2019.100613

Publication details:Yang, Y., Wu, X., Zhou, P., Gou, Z., & Lu, Y. (2019). Towards a cycling-friendly city: An updated review of theassociations between built environment and cycling behaviors (2007–2017). Journal of Transport and Health, 14,[100613]. https://doi.org/10.1016/j.jth.2019.100613

Citing this paperPlease note that where the full-text provided on CityU Scholars is the Post-print version (also known as Accepted AuthorManuscript, Peer-reviewed or Author Final version), it may differ from the Final Published version. When citing, ensure thatyou check and use the publisher's definitive version for pagination and other details.

General rightsCopyright for the publications made accessible via the CityU Scholars portal is retained by the author(s) and/or othercopyright owners and it is a condition of accessing these publications that users recognise and abide by the legalrequirements associated with these rights. Users may not further distribute the material or use it for any profit-making activityor commercial gain.Publisher permissionPermission for previously published items are in accordance with publisher's copyright policies sourced from the SHERPARoMEO database. Links to full text versions (either Published or Post-print) are only available if corresponding publishersallow open access.

Take down policyContact lbscholars@cityu.edu.hk if you believe that this document breaches copyright and provide us with details. We willremove access to the work immediately and investigate your claim.

Download date: 30/06/2020

Contents lists available at ScienceDirect

Journal of Transport & Health

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

Towards a cycling-friendly city: An updated review of theassociations between built environment and cycling behaviors(2007–2017)Yiyang Yanga, Xueying Wub, Peiling Zhoub,∗∗, Zhonghua Gouc, Yi Lua,d,∗

a Department of Architecture and Civil Engineering, City University of Hong Kong, Hong Kong, Chinab School of Architecture, Harbin Institute of Technology, Shenzhen, Chinac School of Engineering and Built Environment, Griffith University, Australiad City University of Hong Kong Shenzhen Research Institute, Shenzhen, China

A R T I C L E I N F O

Keywords:CyclingBuilt environmentPhysical activityUrban designUrban planningHealthy city

A B S T R A C T

Introduction: Cycling behavior has recently attracted great research attention as an importanttype of physical activity and sustainable mode of transportation. In addition, cycling providesother environmental benefits, such as reducing air pollution and traffic congestion. Various builtenvironment factors have been demonstrated to be associated with the popularity of cyclingbehaviors. However, the most recent built environment cycling reviews were conducted nearly10 years ago, and these reviews reached no clear consensus on which built environment factorsare associated with which domain of cycling behaviors. To determine the crucial features of acycling-friendly city, it is therefore necessary to conduct a review based on empirical studies fromthe last decade (2007–2017).Methods: Thirty-nine empirical studies published in peer-reviewed journals between 2007 and2017 were retrieved and reviewed. The results were summarized based on built environmentfactors and four domains of cycling behaviors (transport, commuting, recreation, and general).Weighted elasticity values for built environment factors were calculated to estimate effect sizes.Results: We found consistent associations with large effect sizes between street connectivity andcycling for commuting and transport. The presence of cycling paths and facilities was found to bepositively associated with both commuting cycling and general cycling. However, the effects ofland-use mix, availability of cycling paths to non-residential destinations, and terrain slope oncycling behaviors remained weak. The effects of urban density and other built environmentfactors are mixed.Conclusions: This review has demonstrated that street connectivity and the presence of cyclingpaths and facilities are the two most significant built environment factors that may promotecycling behaviors. With the emergence of advanced measurement methods for both the builtenvironment and cycling behaviors, further studies may overcome current research limitationsand provide robust evidence to support urban planning and public-health practice.

https://doi.org/10.1016/j.jth.2019.100613Received 25 February 2019; Received in revised form 23 July 2019; Accepted 12 August 2019

∗ Corresponding author. Department of Architecture and Civil Engineering, City University of Hong Kong, Hong Kong, China.∗∗ Corresponding author.E-mail addresses: yiyayang-c@cityu.edu.hk (Y. Yang), 954605706@qq.com (X. Wu), lindaplzhou@gmail.com (P. Zhou),

z.gou@griffith.edu.au (Z. Gou), yilu24@cityu.edu.hk (Y. Lu).

Journal of Transport & Health 14 (2019) 100613

Available online 24 August 20192214-1405/ © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).

T

1. Introduction

This review focuses on the impact of built environment factors on cycling behavior, an important form of physical activity.Physical activity has a long-term effect on health and is thus considered by the World Health Organization one of the top fiveinterventions to reduce the prevalence of noncommunicable diseases (Beaglehole et al., 2011). Previous studies have found thatwalking and cycling for transportation purposes can provide substantial net health benefits, after considering the risks of trafficincidents and air pollution exposure (Mueller et al., 2015). Strong evidence demonstrates that cycling behaviors can reduce the risksof premature death, obesity, heart disease, stroke, type II diabetes, metabolic syndrome, colon cancer and breast cancer among adults(Oja et al., 2011; Pucher et al., 2010a,b; Rojas-Rueda et al., 2011; Rojas-Rueda et al., 2013).

Apart from the common health benefits of physical activity in general, cycling behaviors can provide additional health benefits.Cycling particularly enhances cardiorespiratory and metabolic functions (Gotschi et al., 2016; Haennel and Lemire, 2002). Forchildren and adolescents, cycling can effectively improve cardiorespiratory endurance, muscular fitness, bone health, and cardio-vascular and metabolic health (Davison et al., 2008; Voss and Sandercock, 2010). For older adults, cycling can prevent falls, reducedepression, and enhance cognitive functions (James et al., 2013; Rissel and Watkins, 2014).

Furthermore, cycling behaviors provide opportunities to habitually accumulate other types of physical activity, by encouragingpeople to engage in other forms of moderate to vigorous physical activity (MVPA) more frequently and for longer times (Kerr et al.,2016). In addition to its minimal use of fossil fuels and moderate speed, cycling can also be regarded as a sustainable transportationmode that helps to reduce the use of private vehicles and thus mitigates traffic congestion and improves air quality (Li et al., 2011).

Given the health and environmental benefits of cycling, researchers and policymakers seek to understand why bicycle use differsby location and what built-environment interventions can be implemented to increase cycling levels. Previous studies of the asso-ciations between built environments and cycling have demonstrated that several built environment factors are related to differentdomains of cycling behaviors (Day, 2016; Ferdinand et al., 2012; Fraser and Lock, 2011; Heinen et al., 2010; Saelens and Handy,2008; Wang et al., 2016). Cycling for transportation purposes was shown to be positively associated with the presence of dedicatedcycle routes or paths, the separation of cycle tracks from other vehicle paths, high urban density, short travel distances, and theproximity to green space (Fraser and Lock, 2011; Heinen et al., 2010; Saelens et al., 2003a,b). The environmental factors negativelyassociated with cycling for transportation include perceived and objective danger from traffic, long travel distances, steep inclines,and the absence of cycle paths (Fraser and Lock, 2011; Heinen et al., 2010). Reviewers demonstrated that shorter distances and agreater land use mix had a positive association with the proportion of people cycling for commuting, while the evidence of possibleeffects of street connectivity, urban density, the perception of traffic-calming facilities and cycling infrastructure on transportationcycling were mixed (Heinen et al., 2010). For reviews that did not differentiate the domains of cycling, diversely functional andaccessible bicycle parking facilities have been found to increase the likelihood of engagement in cycling behaviors (Heinen et al.,2010; Wang et al., 2016).

Recently, researchers have debated whether a highly walkable built environment also promotes cycling behaviors, or in otherwords, whether walkability equates to bikeability (Muhs and Clifton, 2016). Some researchers have proposed that certain builtenvironment factors, e.g. high urban density, short distance to non-residential destinations, and mixed land use, may promote bothwalking and cycling behaviors (Saelens et al., 2003a,b; Wang et al., 2016), while others concluded that walking and cycling havedifferent associations with built environment characteristics. For example, cycling behaviors were significantly affected by thepresence of cycling infrastructure, such as cycling paths/lanes, and parking facilities, while walking behaviors were not (Buehler andPucher, 2011).

Existing reviews of built environment-cycling associations have a few limitations. First, most of the reviews tended to focus oncycling for transportation purposes. Second, those reviews were conducted nearly 10 years ago (Fraser and Lock, 2011; Heinen et al.,2010; Pucher et al., 2010a,b). Third, the effect sizes for different built environment factors were not estimated. Hence, it is difficult topredict the effectiveness of any planning interventions. Fourth, none of these reviews included any evidence from developingcountries, despite the prevalence of cycling as a major transportation mode in such countries. In China, for example, several largepublic bicycle-sharing programs, e.g. Mobike, provide an affordable and convenient means of shared transportation for short trips.Since the launch of such programs in 2015, the use of bicycles in China has surged, with nearly 19 million Chinese people now usingcycling for their first-mile and last-mile trips (Tsing Hua University Planning and Design Institution and Mobike, 2017). Hence, it isvital to summarize the evidence from recent studies of built environment-cycling associations, as these studies cover a broader scopeof geographical locations with different social and urban contexts.

In this review, we performed a meta-analysis to estimate the effects of individual built environment factors of relatively smallmagnitude on cycling, which are difficult to identify in any single study. Meta-analyses can synthesize outcomes from differentstudies, and calculate and compare the effect sizes across a range of outcomes in different contexts and sample sizes (Borenstein et al.,2009). For example, Ewing and Cervero (Ewing and Cervero, 2010) conducted a meta-analysis of more than 50 studies on asso-ciations between built environment and travel behaviors, with the aim of quantifying the effect sizes and updating earlier relatedwork.

Meta-analyses are useful but also controversial; some limitations are that the results of weaker studies may contaminate those ofstronger ones if the results of such different-strength studies are combined, and that choosing published articles may lead to pub-lication bias, given that results with statistical significance are more likely to be published (Rothstein et al., 2005). Despite thesecaveats, well-designed meta-analyses can effectively summarize existing evidence and guide future research, hence these are widelyused in public health and planning fields (Bartholomew and Ewing, 2009; Cerin et al., 2017; Sharmin and Kamruzzaman, 2017).

In sum, we reviewed studies on the associations between the built environment and cycling behaviors conducted between 2007

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

2

and 2017. We also estimated the effect sizes of those built environment factors based on elasticity values. Finally, we summarized theevidence and discussed potential directions for future research.

2. Method

The data analysis method closely followed that used in Saelens and Handy's review (Saelens and Handy, 2008). These researchersidentified 13 studies published between 2002 and 2006 and another 29 between 2005 and 2006 that focused on the link betweenwalking and the built environment. Day's subsequent review of the correlation between the built environment and physical activityconfirmed the replicability and reliability of the method used by Saelens and Handy (Day, 2016). However, Saelens and Handyfocused on walking in developed countries and Day focused on physical activity in China, this review focuses on built environment-cycling studies in both developed and developing countries published between 2007 and 2017.

2.1. Search terms and criteria

The terms entered into the Web of Science search engine comprised one term from “cycling,” “bicycle,” “bicycling,” “bike,” “bikeuse,” and one term from “built environment,” “physical environment,” “urban form,” “urban”. The search field was limited to articlespublished in peer-reviewed journals in any country and for any population group between 2007 and 2017.

We screened the abstracts of 852 articles in the database and removed those that failed to meet all four criteria below. In the end,our review database consisted of 39 full-text articles, comprising studies which:

1) measured cycling as an individual transport mode or physical activity—instead of being integrated with other modes, such aswalking;

2) examined the association between cycling and certain aspects of objective and/or perceived built environment;3) were published as a research paper in a peer-reviewed journal between 2007 and 2017;4) were written in English.

The workflow of literature review was shown in Fig. 1 (see Fig. 1). Studies that only addressed the impact of social, cultural, andeconomic factors, such as household incomes, and personal attitudes, were excluded. In studies examining multiple factors, theresults for socio-economic and cultural factors were not considered. The built environment factors were urban design, land use, andtransportation system, which are much broader than the scope of physical environment (Handy et al., 2002). Studies that examinedonly the physical environment were also considered in our review database.

2.2. Review scheme

We followed Saelens and Handy's (2008) approach and selected our review scheme from the first 10 papers. The initial aspects weselected to review comprised the sample size, the data source of the built environment, the factors of the built environment examined,and the units of geographical analysis. After these 10 papers had been reviewed, the research scheme was revised for optimization.Four new query aspects were added: study country, domains of cycling, control variables, and whether self-selection was considered.All the papers were reviewed again according to the new optimized research scheme, and the findings were summarized and pre-sented in a detailed table (see Table A.1).

2.3. Summary tables

Table A.1 lists the following information from each paper: study country, sample, source of built environment data, built en-vironmental factors, geographical analyzing units, cycling metrics, controlled variables, whether self-selection was controlled, andresults. The term “cycling metrics” refers to the attributes of cycling behavior, such as duration, frequency, choice of route, andproportion of the cycling population. Self-selection refers to the tendency of people to choose their neighborhood of residence basedon their travel abilities, needs and preferences (Naess, 2009). Hence, the associations observed between built environment andcycling may be explained by personal attitudes and preferences, rather than a true impact of built environment on cycling. Studiesusually attempted to control for residential self-selection, and confounding factors included attitude variables (AT), socioeconomicvariables (SE), weather variables (WE), level of service variables (LE), station variables (ST), and other variables (OT), which referredto the categories and approaches listed in a review conducted by Cao and Mokhtarian (Cao et al., 2009).

Table A.2 elaborates on the details provided in Table A.1. Built environment factors were categorized in terms of measurementmethod (perceived or objective measures) and scale (micro, meso, or macro). Microscale environment is that immediately sur-rounding the respondents, such as streetscape and cycling path design; mesoscale environment refers to the medium-scale en-vironments in which respondents lived, such as neighborhoods or census tracts; and macroscale environment refers to large en-vironments such as cities or countries. We identified 12 built environmental factors from the 39 papers covered in our review study(Fig. 2): density, land use mix, availability of non-residential destinations, accessibility of public transport, street connectivity,availability of green spaces, terrain slope, aesthetics/attractiveness, presence of cycling paths and facilities, cycling safety designfeatures, cycling paths/minor roads length, and composite variables. The classification followed and developed from previous re-views (Day, 2016; Saelens and Handy, 2008; Wang et al., 2016).

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

3

Table 3 is extracted from Table A.2; it underlines the directions of built environment-cycling associations (e.g. expected, un-expected/null). Based on planning theory and previous literature, a greater or better condition of a built environment factor was“expected” to associate with a higher level of cycling behavior (Saelens and Handy, 2008). For terrain slope and slope-related factors,a negative association with cycling was marked as expected, whereas a positive association was regarded as unexpected. For the other11 factors, positive associations were marked as expected, whereas negative associations were marked as unexpected. Any insig-nificant associations were marked as null (Saelens and Handy, 2008).

Based on the data collected on cycling purposes, four domains of cycling were identified: transportation, commuting, recreation,and general cycling, followed by the physical activity domains defined in Day's study (Day, 2016). Cycling for transportation pur-poses was defined as the use of bicycles to travel from one place to another, while cycling for commuting purposes was defined astraveling from home to a place of work or study (Herlihy, 2011). We categorized commuting cycling as an individual domain becausemany empirical studies treat it as a separate outcome, although commuting is usually classified as one subtype of transportation.Cycling for recreation was defined as the voluntary use of a bicycle for satisfaction, pleasure, or creative enrichment (Horner andSwarbrooke, 2005). Studies that did not clearly indicate cycling purposes were classified as general cycling.

2.4. Weighted average elasticity values of built environmental factors

We calculated elasticity values from the 39 studies and then computed the average elasticities weighted by sample size for all 12factors (see Table 4). The elasticity values were unit-free measures of the magnitude of built environment-behavior associations,making it possible to compare the effect size of different built environment factors from various studies with different social and builtenvironment contexts and different estimating techniques for both built environmental factors and cycling behaviors. Elasticity isdefined as the percentage change in a variable associated with a 1% increase in another interest variables, and has been widely usedin recent sensitivity analyses, especially of the built environment and in travel mode studies (Munshi, 2016; B. D. Sun et al., 2017;Zhang, 2004). For instance, if a 1% increase in population density is associated with a 0.3% increase in cycling time, then theelasticity value will be 0.3.

Fig. 1. Workflow of the literature review.

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

4

Fig. 2. The strengths of associations between built environmental factors and different cycling purpose. Strong associations (the number of expectedresults was greater than twice the sum of unexpected and null results) are shown as dark blue lines. Emerging associations (where the number ofexpected results was more than the sum of unexpected and null results but less than twice of the sum) are shown as light blue lines. (For inter-pretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Table 1Elasticity calculation formulas.

Regression model Elasticity Formula

Linear × XY¯¯

Log-logLogistic × × ( )X 1 Y

n¯

Note: β is the regression coefficient on the built environmentfactors in a specified study. Y refers to the mean value of thecycling behavior in a specified study, and X refers to the meanvalue of the built environment factors. Y

n¯ is the mean estimated

probability of occurrence.

Table 2Number of studies available for the computation of weighted average elasticity by sample size. Weighted elasticity values were calculated if therewere at least three candidate studies (shown in the bolded font).

Number of Available Studies

Transport Commuting Recreation General

Density 4 0 0 0Land use mix 3 0 0 0Availability of non-residential destinations 1 0 0 0Accessibility of public transport 0 0 0 0Street connectivity 4 4 0 0Availability of green spaces 1 0 0 3Terrain slope 0 0 0 3Aesthetics/attractiveness 1 0 0 0Presence of paths and facilities 5 3 0 3Cycling safety design features 6 0 0 0Cycling paths/minor roads length 0 0 0 0Composite variables 0 0 0 0

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

5

The calculations of elasticities for individual studies were conducted in one of two ways (Ewing and Cervero, 2010). We either: (1)found elasticity values reported in papers; or (2) calculated a value from regression coefficients and the mean values of dependentand independent variables. For method (2), we used the formulas in Table 1 to compute elasticity values, depending on the types ofregression method used to estimate coefficient values. Using formulas to calculate elasticity from mean values may result in an errorin estimated results because the difference between the mean and individual values of elasticities can be significant. Train mentionedthat “the probability evaluated at the average utility underestimates the average probability when the individuals’ choice prob-abilities are low and overestimates when they are high” (Train, 1986, p. 42). When the elasticity is calculated over a curve re-lationship, it refers to an arc elasticity. However, due to the limited accessibility of raw data, we nevertheless used the mean values tocalculate the elasticities quoted in this review.

The weighted average elasticity of each built environment factor was calculated according to three conditions: 1) studies reportedsignificant results of a built environmental factor; 2) studies reported descriptive statistics of dependent and independent variables,which are necessary to compute elasticity value; and 3) at least three studies were available. Number of candidate studies used tocalculate weighted average elasticity values are shown in Table 2. Ten of the 39 studies included in this review for computations ofweighted average elasticity used objective measurement of the built environment factors. Only one study (Ma and Dill, 2015) thatused perceived measures to compute weighted average elasticity was included, while other studies using perceived measures did notreport adequate descriptive statistics for elasticity calculation.

3. Results

3.1. Study sample

We reviewed 39 studies published between 2007 and 2017 focusing on associations between built environment and cycling. Moststudies (33 out of 39) were conducted in developed countries; three studies were conducted in developing countries; and three studieswere conducted in multiple countries including both developing and developed countries. The top three studied countries were theUnited States (12 studies), Belgium (10), and Australia (9). There were also studies conducted in developing countries such as Brazil

Table 3Strength of association grouped according to environmental factors and domains of cycling and expected evidence (+), unexpected evidence (−)and null (0) results. One study may have multiple results. Strong evidence (where the number of expected results was more than twice the sum of theunexpected and null results) is shaded in dark blue; emerging evidence (where the number of expected results was more than the sum of unexpectedand null results but less than twice of the sum) is shaded in light blue.

Table 4Weighted average elasticity values of built environment factors.

Elasticity value

Transport Commuting Recreation General

Density < 0.01 – – –Land use mix 0.09 – – –Availability of non-residential destinations – – – –Accessibility of public transit – – – –Street connectivity 0.08 0.39 – –Availability of green spaces – – – 0.27Terrain slope – – – 0.26a

Aesthetics/attractiveness – – – –Availability of cycling facilities/paths 0.07 0.28 –Cycling safety design features 0.04 – – –Cycling paths/minor road lengths – – – –Composite variables – – – –

a Sign was reversed for terrain slope.

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

6

(3), China (2), Colombia (3), and Mexico (2). The criteria used to define a country as “developed” and “developing” followed theStandard Country or Area Codes for Statistical Use (M49) published by the United Nations (United Nations, 1999).

Five studies focused on children or adolescents, and four focused on older adults. In terms of sample size, over 60% of studiesrecruited fewer than 5000 participants, while 14 of 39 studies recruited fewer than 1000 participants. Studies conducted in multiplecities usually recruited more than 10,000 participants.

3.2. Built environment

Most of the studies (20 out of 39) focused on the impact of the built environment at mesoscale (at neighborhood level) andcollected built environmental factors within a buffer of a respondent's residential locations.

3.2.1. Perceived or objective measuresIn terms of the measurement of built environment factors, objective measures were used in 26 of the 39 studies; perceived

measures were used in 22 studies; and perceived and objective measures were simultaneously used in 9 studies. While one papersuggested that both perceived and objective measures had distinct effects on cycling behaviors (Ma and Dill, 2015), other studiesargued that the effects of perceived measures were slightly more reliable than those of objective measures (Orstad et al., 2017).

3.2.2. Data collecting methodFor objective measures, Geographic Information System (GIS) technology, combined with census data and open source urban

land-use data, has been widely used to identify objective built environment measures. Recent studies have also used advancedmethods to improve the accuracy of measurement, such as the Normalized Difference Vegetation Index (NDVI), a greenness indexbased on satellite remote sensing images (Tilt et al., 2007).

For perceived measures, interviews and questionnaires were the most common data-collection methods. In particular, theNeighborhood Environment Walkability Scale (Cerin et al., 2006; Saelens et al., 2003a,b; Sallis, 2002) was the most widely usedquestionnaire to measure mesoscale built environments; various versions of this questionnaire have been developed and validated indifferent study areas. Besides standard questionnaires or interviews, one recent study has exploited the opportunities of using si-mulated photos to survey perceptions of proposed built environment changes (Ghekiere et al., 2015).

3.2.3. Built environment factorsIn our review database, cycling safety design features (addressed in 32 of the 39 studies) were the most widely studied aspect of

the built environment, followed by street connectivity (26 of 39 studies), presence of cycling paths and facilities (23 of 39 studies),and land use mix (22 of 39 studies).

3.3. Cycling

3.3.1. Domains of cyclingWe found that 23 of the 39 studies examined cycling for transportation. Nine studies examined cycling for commuting and eight

for cycling for recreational purposes, respectively. Ten studies focused on general cycling behaviors.

3.3.2. Data collection methodThe most common method of collecting cycling data, such as cycling time and frequency, was by questionnaires. Fourteen studies

used the International Physical Activity Questionnaire (IPAQ) or a modified IPAQ, whereas many other studies developed their ownquestionnaires. The participants were asked to report the duration or frequency of their cycling activity during a given period, and/ortheir willingness to cycle for different purposes.

In studies focusing on microscale built environment, some researchers used a mapping method, e.g. retracing a cycling route on amap to collect participants' cycling route choice, which is an important indicator of participants’ preference of cycling environment(Snizek et al., 2013). In studies focusing on macroscale built environment attributes, such as at city level and/or across countries, theproportion of the cycling population and the count of cyclists in specific places were also widely used to measure cycling behaviors(Buehler and Pucher, 2011).

3.4. Associations between cycling and the built environment

A built environment factor was identified as strong evidence if the number of expected results was more than twice the sum of theunexpected and null results; a factor was identified as an emerging evidence if the number of expected results was more than the sumof unexpected and null results but less than twice of the sum. In terms of transportation cycling, there was strong evidence for anassociation with street connectivity (11 expected results vs. 5 unexpected/null), and availability of non-residential destinations (5 vs.2), and emerging evidence for an association with a composite index (2 vs. 1) (Table 3). However, the evidence for an associationwith other built environment factors was weak (i.e. the number of expected results was less than or equal to the number of

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

7

unexpected results).For commuting cycling behavior, there was strong evidence of an association with street connectivity (3 vs. 1) and presence of cycling

routes/paths (5 vs. 2), and emerging evidence for an association with land use mix (2 vs. 1) and availability of green spaces (2 vs. 1).The results regarding recreational cycling behavior were less clear, partly because there were fewer of these. We did observe

distinct evidence in these studies for associations between recreational cycling and any built environmental factors, although therewas inconclusive evidence for an association with street connectivity, given the equal number of expected and unexpected/nullresults (4 vs. 4).

For general cycling behavior, strong evidence was found of an association with the presence of cycling routes/paths (5 vs. 0), openspace and green space (3 vs. 1), and aesthetics and attractiveness (3 vs. 1); and emerging evidence was found for an association withterrain slope (3 vs. 2), and cycling safety design features (4 vs. 3).

3.5. Elasticity values of built environmental factors

We calculated elasticity values from individual studies in our review database. Table 3 shows the weighted average elasticityvalues of 12 built environmental factors on different domains of cycling. For transport cycling, land use mix had the largest elasticityvalue of 0.09, which meant that a 1% increase in land use mix was associated with a 0.09% increase in cycling for transport. Streetconnectivity had the second largest elasticity value of 0.08, while other factors had relatively small elasticity values below 0.05. Forcommuting cycling, both street connectivity (0.39) and availability of cycling facilities/paths (0.28) have relatively large elasticityvalues.

4. Discussion

In this study, we conducted an updated review of the associations between the built environment and cycling behaviors based on39 empirical studies conducted in 2007–2017. The results showed that different domains of cycling behaviors (transport, commuting,recreation, and general) may have different associations with built environment factors. Consistent and positive associations werefound between street connectivity and cycling for transportation and commuting purposes, which suggests that street connectivitymay be a fundamental requirement for cycling behaviors. The presence of cycling paths and facilities and open/green spaces wasassociated with both commuting and general cycling behaviors, which suggests that these two factors are more important forcommuting cycling than other domains.

The land use mix, availability of paths to non-residential destinations, and composite variables are only associated with certaindomains of cycling, and show less evidence of an association with general cycling. By contrast, low slope and cycling safety are onlyassociated with general cycling but show less evidence of association with any specific domains. However, it is still difficult to drawdefinitive conclusions about these specific associations given the limited number of available studies. A future review of a sufficientnumber of studies, especially those simultaneously considering different domains of cycling behaviors, is warranted; such a reviewmay unequivocally demonstrate the relationship of particular built environment characteristics with particular cycling behaviors(Ferdinand et al., 2012).

4.1. Comparison with previous reviews

Previous reviews suggested that land use mix and provision of bicycle parking facilities may increase the likelihood of engage-ment in general cycling (Saelens et al., 2003a,b; Wang et al., 2016). We found strong evidence in results from studies during2007–2017 for associations between general cycling and availability of green spaces, terrain slope and availability of cycling fa-cilities/path. All three factors had relatively large elasticity values.

Previous reviews found significant associations between transport cycling behaviors and the presence of cycle routes/paths, highpopulation density, availability of non-residential destinations, and availability of green space, and terrain slope (Fraser and Lock,2011; Heinen et al., 2010; Saelens et al., 2003a,b). The results of the 2007–2017 studies regarding transport cycling found strongassociations with street connectivity, and emerging associations for the influence of availability of non-residential destinations andcomposite factors upon cycling behaviors, although the elasticity values for these factors was relatively small.

For commuting cycling, previous reviews document strong associations with land use mix (Heinen et al., 2010). This review alsofound evidence of consistent associations between commuting cycling and land use mix, and constant associations with the avail-ability of green space, street connectivity, and presence of cycling paths and facilities. The latter two factors also had relatively largeelasticity values. This indicates the great importance of land use mix and green spaces and the increasing importance of streetconnectivity and cycling paths/facilities in promoting commuting cycling.

However, the recreational cycling findings of the 2007–2017 studies showed no clear association with most of the environmentalfactors. The unexpected/null results outnumbered expected results for any environmental factors, which echoed findings fromprevious reviews (Day, 2016; Saelens et al., 2003a,b).

In contrast with the findings of previous reviews, our review did not reveal a clear association between cycling behaviors andurban density. This contrasting result may be explained by increased ranges of urban density arising from a broader scope of

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

8

geographical locations covered in this review. That is, this review included studies conducted in both developed and developingcountries, while previous reviews included only studies conducted in developed countries (Christiansen et al., 2016). In addition,cities classified as “high density” in China and South America may be several times denser than “high-density” cities in Europe, NorthAmerica and Australia. Indeed, researchers have recently indicated that different ranges of urban density may account for theinconsistent influence of urban density on physical activity (Gomez et al., 2010; Lu et al., 2017; Su et al., 2014; Szeto et al., 2017).

4.2. Objective measures vs. perceived measures

In addition to more evidence of associations between the built environment and cycling, studies in the past decade have in-corporated significant methodological developments in addressing some limitations raised in previous reviews (Christiansen et al.,2016; Cole-Hunter et al., 2015; Day, 2016; Fraser and Lock, 2011; Ma and Dill, 2015). One such development was to explore therelationship between the objective and perceived built environment. The objective measures of built environment factors are oftenobtained from existing GIS data or field audits, while perceived measures are often collected based on an individual's perception ofbuilt environment features (Brownson et al., 2009). The results obtained from objective measures can be repeated and hence are morereliable than results from perceived measures. However, perceived measures reflecting individual exposure and perceptions of thebuilt environment, and thus may complement and inform studies based on objective measures (Ma and Dill, 2015). For example,safety and aesthetics are typically assessed by subjective measures, but objective measurement of these factors is warranted in furtherstudies.

4.3. Bikeability and walkability

As mentioned at the beginning of this review, researchers have debated whether built environment factors promoting walkingalso promote cycling (Muhs and Clifton, 2016). Previous reviews suggested that higher-density neighborhoods with more diverseland use and greater street connectivity may promote walking behaviors (Saelens and Handy, 2008; Wang et al., 2016). However, wedid not find a close relationship between walkability and bikeability, as only street connectivity was positively associated withcycling. Recent empirical studies investigating walking and cycling simultaneously have also argued that walking and cycling areaffected by different built environment factors, and to a different extents (Muhs and Clifton, 2016; Ton et al., 2019).

4.4. Planning and policy applications

This meta-analysis provides the elasticity values of various built environment measures on cycling behaviors from pooled samples.These values may be useful for urban planners and policymakers for providing rough forecasts on baseline cycling behaviors for newurban design projects, or of potential changes in cycling behaviors after design interventions in existing urban areas. Previous reviewshave already suggested the utilization of elasticities, and the new elasticities calculated in this review could also be used in similarsituations (Cervero, 2006; DKS Associates, 2007; Johnston, 2004; Walters et al., 2000). Furthermore, a potential application ofelasticities may be to estimate health outcomes. Knowledge of regular cycling behaviors, as a common form of physical activity, iscritical to predicting health conditions. Therefore, a potential increase in cycling can be used to project the potential health impacts ofurban design interventions. Similarly, it can also be used to estimate the reduction of greenhouse gas emissions or air pollution,because cycling can reduce private vehicle use.

4.5. Limitations

There are three limitations to this review. First, because we only reviewed articles written in English, the results of researchwritten in other languages were not considered. However, this review has covered a broader range of geographical locations, in-cluding both developed and developing countries, whereas previous reviews largely focused on developed countries (Christiansenet al., 2016).

Second, only five studies covered in this review involved children or adolescents. Being active during childhood has been found todecrease several health risk factors and enhance quality of life in later years (Aires et al., 2011; Bailey et al., 2012; Maki-Opas, deMunter, Maas, den Hertog and Kunst, 2014). Future studies should thus pay more attention to children's and adolescents' cyclingbehaviors, due to the increasing trend of obesity among young people (Cai et al., 2017).

Third, it was still a challenge to assess causality from the 2007–2017 studies. Most recent studies have used cross-sectionalresearch design, making it difficult to determine whether the built environment alters residents’ cycling behaviors or if cyclists chooseto live in areas with cycling-supporting built environment characteristics. To address this limitation, natural experiment researchdesigns should be implemented to examine causal relationships. For example, researchers from Perth made an attempt to address thisissue by conducting a RESIDE project, a longitudinal natural experiment to compare cycling behaviors of people before and aftermoving into a new residence (Badland et al., 2013). The results demonstrated that changes in both objective and self-reported builtenvironment characteristics were associated with changes in transport cycling and recreational cycling. This longitudinal

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

9

experimental design enabled the researchers to study the effects of neighborhood design on cycling behaviors while controlling forindividual attitudes toward cycling and other health-related behaviors.

4.6. Future directions

As objective measurements of built environments can avoid potential bias in self-reported data, big data in conjunction withmachine learning techniques offer unprecedented opportunities for objectively measuring fine-grained built environment char-acteristics. Open source data have been widely used in recent urban studies (Cole-Hunter et al., 2015; Hino et al., 2014; Y. R. Sunet al., 2017). OpenStreetMap, Points of Interest, and Google Street View (GSV) provide plenty of opportunities to quantify builtenvironment metrics, such as street networks and streetscapes (Lu, 2018; Y. R. Sun et al., 2017; Wang et al., 2019a, 2019b). Computervision is one of the most popular machine learning applications and is used mainly for object recognition. GSV and other imagedatabases can help researchers to develop image recognition methods to measure the built environment. SegNet is a notable ap-plication developed by the University of Cambridge: it is a fully convolutional deep neural network architecture for semantic pixel-wise segmentation, trained to classify urban street images into 12 categories, such as ‘sky’, ‘building’, ‘road’, ‘pavement’, and ‘tree’,enabling much more reliable assessment of the percentage of streetscape design elements (Badrinarayanan et al., 2017). Researchershave also developed several objective variables to measure urban greenness, such as NDVI, measured using high-solution satelliteremote images, and GSV eye-level greenness (Cole-Hunter et al., 2015; Lu et al., 2018). The advance of machine learning methodsmay significantly improve the accuracy of estimating a person's environment exposure (e.g. eye-level greenness, building density),and is more time- and cost-effective than field audits.

As perceived data can reflect individual exposure and perceptions of the built environment, it would be better for such data to bemeasured objectively. To this end, portable sensing technologies, notably GPS technology, have enhanced the capacity for recording,measuring, and analysis of human behaviors, and use of these would thereby avoid the bias of self-reported built-environment data.The emergence of advanced equipment, such as eye tracking or mobile electroencephalography, may enable the assessment andrecording of the psychological impact of environments on individuals, and hence provide opportunities for researchers to develop anenhanced understanding of the relationship between the perceived environment and cycling behaviors (Mavros et al., 2016).

Such technological advancement may also reduce the cost of collecting detailed behavioral data. Crowd-sourced GPS data andphysical-activity apps are emerging methods of collecting large samples of cycling data (B. D. Sun et al., 2017). However, the extentto which app users represent the total population of cyclists in a given city remains unclear (Selala and Musakwa, 2016). In China,public bicycle-sharing programs have attracted millions of people back to cycling. These shared bicycles are usually equipped withGPS technology and the data of the routes on which these bicycles are taken are collected (Tsing Hua University Planning and DesignInstitution and Mobike, 2017). Such data may help researchers use large sample sizes to further explore the association between thebuilt environment and cycling behaviors.

Most of the 2007–2017 studies were still conducted in cities of Europe, North America, and Australia, although some emergingstudies were conducted in cities of developing countries, such as Brazil, China, Colombia, and Mexico. Future studies should pay moreattention to the cities of Asia and South America, where governments are investing in cycling infrastructure and cycling is becomingmore and more popular. Furthermore, given the fact that cycling behaviors may be influenced by local culture and vehicle ownership,the studies conducted in a broader range of geographical locations may improve our understanding of the relationships between thebuilt environment and cycling in diverse urban and cultural contexts.

This review focused on the unique characteristics of cycling, and studies with the same research framework will provide betterguidance for policymakers, urban planners and designers in creating a cycling-friendly urban environment, including more specificrecommendations for different demographic groups and different cities.

5. Conclusion

This review summarizes the recent evidence (2007–2017) of the relationship between cycling and various built environmentcharacteristics. A positive and consistent correlation between street connectivity and cycling for commuting and other transportationpurposes was found. In addition, the presence of cycling paths and facilities was found to be positively associated with both com-muting cycling and general cycling. There were weak or mixed associations with other built environment factors, such as land usemix and density. With the emergence of advanced measurement methods, future studies should overcome current limitations andsupport improved urban planning and public-health practice to increase uptake of cycling.

Funding

The work described in this paper was fully supported by grants from the National Natural Science Foundation of China (ProjectNo.51578474, 51778552) and the Research Grants Council of the Hong Kong Special Administrative Region, China (Project No.CityU11666716).

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

10

App

endixA

Tabl

eA.1

Char

acte

riza

tion

ofstud

iesby

coun

try,

sam

ple,

built

envi

ronm

enta

ldat

aso

urce

,bui

lten

viro

nmen

talf

acto

rs,a

naly

zed

geog

raph

icun

it,cy

clin

gm

etrics

,con

trol

led

variab

les,

self-

sele

ctio

nco

ntro

lled

orno

t,an

dm

ajor

resu

lts.

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

1Ghe

kier

eet

al.

(201

5)

Belg

ium

305

fifth

and

sixt

hgr

ade

child

ren

and

thei

rpa

rent

sfrom

twel

vera

ndom

lyse

-le

cted

prim

ary

scho

ols

Pano

ram

icco

lor

phot

ogra

phs

1.Ev

enne

ssof

cycl

epa

th2.

Deg

ree

ofse

para

tion

from

mot

oriz

edtraffi

c3.

Spee

dlim

itfo

rm

otor

-iz

edtraffi

c

Indi

vidu

alre

spon

-de

ntRo

ute

choi

cefrom

child

ren

and

thei

rpa

rent

sre

spec

-tiv

ely

SEYe

sPr

efer

ence

ofch

ildre

n:1.

Even

ness

ofth

ecy

cle

path

2.Lo

wer

spee

dlim

ita-

tion.

Pref

eren

ceof

pare

nts:

1.Lo

wer

spee

dlim

ita-

tion

2.Deg

ree

ofse

para

tion

with

mot

oriz

edtraffi

c2

Titz

eet

al.

(200

8)

Aus

tria

905

inha

bita

ntsag

ed15

–60

year

sfrom

Gra

zIn

terv

iew

1.Attra

ctiv

enes

sof

cy-

clin

gco

nditi

ons

2.La

ndus

em

ix-d

iver

sity

ofus

es3.

Safe

tyfrom

traffi

c4.

Bike

lane

conn

ectiv

ity5.

Pres

ence

ofstee

pel

e-va

tion

6.Pr

esen

ceof

stre

et-

light

sdu

ring

nigh

t7.

Pres

ence

ofside

wal

k8.

Hom

e:bi

cycl

epa

rkin

g9.

Des

tinat

ion:

bicy

cle

park

ing

Indi

vidu

alre

spon

-de

ntSh

are

ofcy

clin

gSE

/OT

No

Lane

conn

ectiv

ity,t

hepr

esen

ceof

astee

pel

e-va

tion

wer

epo

sitiv

ely

asso

ciat

edw

ithbi

cy-

clin

gfo

rtran

spor

tatio

n.

3Sy

lvia

Titz

eet

al.(

2010

)

Aus

tral

ia18

13pa

rtic

ipan

tsag

era

nge

18–7

8ye

arsfrom

Perth

Nei

ghbo

rhoo

dEn

viro

nmen

tW

alka

bilit

ySc

ale

1.Acc

essto

serv

ices

(lan

dus

em

ix-a

cces

s)2.

Nei

ghbo

rhoo

dsu

r-ro

undi

ngsgr

een

and

attrac

tive

3.Bi

cycl

e/w

alki

ngpa

ths

acce

ssib

le4.

Num

berof

destin

a-tio

nsfo

rtran

spor

tcy-

clin

g5.

Num

berof

destin

a-tio

nsfo

rre

crea

tiona

lcy

clin

g6.

Man

ytraffi

cslow

ing

devi

ces

7.Pr

esen

ceof

man

y4-

way

inte

rsec

tions

Nei

ghbo

rhoo

d1.

Tim

esof

cy-

clin

gfo

rtran

s-po

rtpe

rw

eek

2.Ti

mes

ofcy

-cl

ing

forre

-cr

eatio

npe

rw

eek

SE/A

TYe

s1.

Leaf

yan

dat

trac

tive

neig

hbor

hood

s,ac

-ce

ssto

bicy

cle/

wal

king

path

s,th

epr

esen

ceof

traffi

cslow

ing

devi

cesan

dha

ving

man

y4-

way

stre

etin

ters

ectio

nsw

ere

positiv

ely

as-

soci

ated

with

cy-

clin

gfo

rtran

spor

t.2.

The

perc

eptio

nof

al-

tern

ativ

ero

utes

for

gettin

gfrom

plac

e-to

-pl

ace

sign

ifica

ntly

in-

crea

sed

the

odds

ofcy

clin

gfo

rre

crea

tion

com

pare

dw

ithlo

w

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

11

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

8.Pr

esen

ceof

man

yal

-te

rnat

ive

rout

es9.

Maj

orba

rrie

rsin

loca

lar

ea10

.Abs

ence

ofcu

l-de-

sacs

perc

eive

dpr

eval

ence

ofal

tern

ativ

ero

utes

.

4M

aan

dDill

(201

5)

Am

eric

a90

2ad

ults

from

Portla

ndSu

rvey

,Re

gion

alLa

ndIn

form

atio

nSy

stem

(RLI

S)from

Portla

ndM

etro

,Re

gion

'stran

spor

ta-

tion

and

land

use

plan

ning

agen

cy,

U.S

.Cen

susBu

reau

1.Pe

rcep

tion

ofoff

-stre

etbi

kepa

ths

2.Pe

rcep

tion

ofbi

kela

nes

3.Pe

rcep

tion

ofqu

iet

stre

etsea

syfo

rbi

ke4.

Perc

eptio

nof

man

ybi

kede

stin

atio

nsne

arby

5.M

ilesof

off-stree

tbik

epa

thw

ithin

1/2-

mile

buffe

r6.

Mile

sof

bike

lane

with

in1/

2-m

ilebu

ffer

7.M

ilesof

min

orstre

etw

ithin

1/2-

mile

buffe

r8.

#Re

tail

jobs

(000

)w

ithin

1/2-

mile

buffe

r

Nei

ghbo

rhoo

dNum

berof

days

ofcy

clin

gfo

rco

m-

mut

ing

ortran

s-po

rtat

ion

over

the

past

mon

th

SE/O

T/AT

Yes

1.Re

spon

dent

wou

ldha

vebe

enm

ore

likel

yto

bicy

cle

for

tran

spor

tatio

nif

they

perc

eive

dth

atth

ere

are

man

ybi

keab

lede

stin

a-tio

nsne

arth

eir

hom

e.2.

Resp

onde

ntw

ould

have

been

mor

elik

ely

tobi

cycl

efo

rtran

s-po

rtat

ion

ifm

ilesof

min

orstre

etsis

long

er.

3.Pe

rcei

ved

and

obje

c-tiv

ebu

ilten

viro

nmen

tha

vein

depe

nden

tef

-fe

ctson

the

bicy

clin

gpr

open

sity

.5

Ow

enet

al.

(201

0)

Aus

tral

iaan

dBe

lgiu

m21

59pa

rtic

ipan

tsag

ed20

–65

year

sfrom

Ade

laid

e(A

ustral

ia),

and

382

partic

ipan

tsag

ed18

–65

year

sfrom

Ghe

nt(B

elgi

um)

Ade

laid

e:Ce

nsus

Colle

ctio

nDistric

ts(C

Ds)

via

GIS

data

base

Ghe

nt:

Surv

ey,

Nei

ghbo

rhoo

dEn

viro

nmen

tW

alka

bilit

ySc

ale

Que

stio

nnai

re

Ade

laid

e:1.

Dw

ellin

gde

nsity

2.St

reet

conn

ectiv

ity3.

Land

use

mix

4.Net

reta

ilar

eara

tioto

cond

ucta

wal

kabi

lity

inde

x.Ghe

nt:

1.Re

side

ntia

lden

sity

2.La

ndus

em

ix(d

is-

tanc

eto

faci

litie

s)3.

Conn

ectiv

ity

Nei

ghbo

rhoo

dNum

berda

ysof

cycl

ing

fortran

s-po

rtin

aw

eek

(at

leas

t10

min

)

SENo

1.Ade

laid

e,pe

ople

livin

gin

ahi

gher

-w

alka

ble

com

mu-

nity

had

82%

high

erod

dsof

regu

larbi

-cy

cle

use

fortran

s-po

rt.

2Ghe

nt,p

eopl

eliv

ing

inth

ehi

gher

and

high

estw

alka

bilit

yne

ighb

orho

odsha

dap

prox

imat

ely

a2.

5tim

eshi

gher

likel

i-ho

odto

bicy

cle

for

tran

spor

t6

Win

ters

etal

.(2

016)

Cana

daan

dAm

eric

a56

64ce

nsus

trac

tsin

24US

and

Cana

dian

citie

sW

alk

Scor

e1.

Bike

Lane

Scor

e2.

Hill

Scor

e3.

Des

tinat

ions

and

Conn

ectiv

itySc

ore

1.Nei

ghbo

rhoo

d2.

City

Shar

eof

cycl

ing

for

com

mut

ing

OT

No

Att

hece

nsus

trac

tlev

el,

ate

n-un

itin

crea

sein

Bike

Scor

ew

asas

so-

ciat

edw

itha

0.5%

in-

crea

sein

the

prop

ortio

nof

popu

latio

ncy

clin

gto

wor

k

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

12

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

7Hin

oet

al.

(201

4)

Braz

il12

06ad

ults

aged

abov

e16

year

sfrom

Curitib

aGIS

1.Bu

sstop

num

ber

2.BR

Ttu

bestat

ions

num

ber

3.Tr

affic

safe

tynu

mbe

r4.

Entrop

y(lan

dus

ehe

tero

gene

ity)

5.Re

side

ntia

lare

apr

o-po

rtio

n6.

Com

mer

cial

area

pro-

portio

n7.

Stre

etde

nsity

8.Num

berof

bloc

ks9.

Ave

rage

leng

thof

the

stre

ets

10.Dea

d-en

dstre

etspr

o-po

rtio

n11

.St

reet

inte

rsec

tions

(≥4

way

)pr

opor

tion.

12.Sl

ope

13.Bi

kepa

thde

nsity

14.Dista

nce

tone

ares

tbu

sstop

15.Dista

nce

tone

ares

tBR

Ttu

bestat

ion

16.Dista

nce

tone

ares

tbi

kepa

th

Nei

ghbo

rhoo

dDur

atio

ntim

eof

cycl

ing

fortran

s-po

rtat

ion

perw

eek

(min

s)

SENo

Gre

ater

num

berof

traffi

clig

hts,

and

high

erla

ndus

em

ixw

ere

in-

vers

ely

asso

ciat

edw

ithcy

clin

g.

8Ch

enet

al.

(201

7)

Am

eric

abi

cycl

eco

untda

taco

llec-

tion

in5

year

sfrom

Seat

tle

Seat

tleDep

artm

ent

ofTr

ansp

orta

tion

and

Puge

tSo

und

Regi

onal

Coun

cil

ina

1-m

ile,0

.5-m

ile,0

.25-

mile

buffe

r,1.

The

sum

ofbi

kela

nele

ngth

,in

mile

s;2.

The

perc

entof

stee

par

easin

buffe

rs,i

n%

;3.

The

perc

entof

wat

ersp

aces

inbu

ffers

,in

%;

4.Th

epe

rcen

tof

com

-m

erci

alan

dm

ixed

land

use

inbu

ffers

,in

%;

5.Th

epe

rcen

tof

resi-

dent

iall

ands

inbu

f-fe

rs,i

n%

;6.

The

entrop

yof

six

type

sofl

and

use,

in%

;7.

The

perc

entof

gree

nsp

aces

and

park

sin

buffe

rs,i

n%

;

Nei

ghbo

rhoo

dCo

unts

ofbi

cycl

eat

inte

rsec

tions

WE/

OT

No

1.Bi

cycl

eco

unts

are

grea

terin

zone

sw

ithm

ore

mix

edla

ndus

e,a

high

erpe

rcen

tage

ofw

ater

bodi

es,o

ra

grea

ter

perc

enta

geof

wor

k-pl

aces

;2.

The

incr

emen

tsof

bi-

cycl

ein

fras

truc

ture

ispo

sitiv

ely

asso

ciat

edw

ithth

ein

crea

seof

bicy

cle

volu

me;

3.Bi

cycl

eco

unts

are

few

erin

stee

par

eas

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

13

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

8.Th

epe

rcen

tof

office

san

dgo

vern

men

tsin

buffe

rs,i

n%

;9.

The

num

berof

em-

ploy

men

tspe

rsq

uare

mile

10.Th

enu

mbe

rof

hous

e-ho

ldspe

rsq

uare

mile

9W

inte

rset

al.

(201

0)

Cana

da19

02ad

ults

aged

abov

e19

year

sfrom

Met

roVa

ncou

verre

gion

1.Ce

nsus

,2.

Aca

dem

icre

-se

arch

proj

ects,

3.Pr

oper

tyta

xas

-se

ssm

ent

auth

ority

4.Re

gion

altran

sit

auth

ority

5.GIS

1.%

ofla

ndar

eaw

ithgr

een

cove

r;2.

Variat

ion

inel

evat

ion

(hill

ines

s);

3.%

road

segm

ents

with

slop

e95

%(s

teep

hills

);4.

Ratio

of4-

way

inte

r-se

ctio

nsto

alli

nter

-se

ctio

ns;

5.%

ofro

adne

twor

kth

atis

Hig

hway

;6.

%of

road

netw

ork

that

isArter

ialr

oad;

7.%

ofro

adne

twor

kth

atis

loca

lroa

d;8.

%of

road

netw

ork

that

isoff

-stree

tpat

h;9.

%of

road

netw

ork

that

isDes

igna

ted

bike

rout

e;10

.Pr

esen

ceof

Traffi

cca

lmin

gfe

atur

es;

11.Pr

esen

ceof

Road

mar

king

sor

sign

age

forcy

clists;

12.Pr

esen

ceof

cros

sing

sw

ithcy

clist-a

ctiv

ated

traffi

clig

hts;

13.Po

pula

tion

perhe

c-ta

re;

14.La

ndus

em

ix;

15.%

ofla

ndus

eth

atis

sing

lefa

mily

reside

n-tia

l;16

.%

ofla

ndus

eth

atis

mul

tiple

fam

ilyre

si-

dent

ial;

Zone

ofro

ute,

orig

inan

dde

stin

a-tio

n

1.Th

em

ostf

re-

quen

tnon

-re-

crea

tiona

lbi-

cycl

etrip

(if

any)

,2.

Any

othe

rno

n-re

crea

tiona

lbi-

cycl

etrip

(if

any)

SENo

Incr

ease

dod

dsof

bicy

-cl

ing

wer

eas

soci

ated

with 1.

Less

hilli

ness

;2.

Hig

herin

ters

ectio

nde

nsity

;3.

Less

high

way

san

dar

-te

rial

s;4.

Pres

ence

ofbi

cycl

esign

age,

traffi

cca

lmin

g,an

dcy

clist-

activ

ated

traffi

clig

hts;

5.M

ore

neig

hbor

hood

com

mer

cial

,edu

ca-

tiona

l,an

din

dustrial

land

uses

;6.

Gre

ater

land

use

mix

;7.

Hig

herpo

pula

tion

dens

ity.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

14

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

17.%

ofla

ndus

eth

atis

Com

mer

cial

;18

.%

ofla

ndus

eth

atis

Educ

atio

nal;

19.%

ofla

ndus

eth

atis

Ente

rtai

nmen

t;20

.%

ofla

ndus

eth

atis

Indu

strial

;21

.%

ofla

ndus

eth

atis

Offi

ce;

22.%

ofla

ndus

eth

atis

Park

.10

Mer

tens

etal

.(2

017)

Belg

ium

,Fra

nce,

Hun

gary

,The

Net

herlan

dsan

dth

eUni

ted

King

dom

6037

indi

vidu

alsag

edab

ove

18ye

arsfrom

Ghe

ntre

gion

(Bel

gium

),Pa

risre

gion

(Fra

nce)

,gr

eate

rBu

dape

st(H

unga

ry),

Rand

stad

re-

gion

(The

Net

herlan

ds)

and

Gre

ater

Lond

on(the

Uni

ted

King

dom

)

SPOTL

IGHT

Virtua

lAud

itTo

ol(s

-VAT)

1.Pr

esen

ceof

traffi

cca

lmin

gfe

atur

es(s

uch

assp

eed

hum

ps,

traffi

cisla

nd,r

ound

-ab

outs

ortraffi

clig

hts)

;2.

Spee

dlim

it≤

30km

/h;

3.Abs

ence

ofbi

cycl

ela

nes;

4.Ca

rsth

atfo

rman

ob-

stac

leon

the

road

;5.

Pres

ence

ofgr

een

and

wat

erar

eas;

6.Pr

esen

ceof

tree

s;7.

Pres

ence

oflit

ter.

Nei

ghbo

rhoo

din

the

last

7da

ys:

1.Num

bero

fday

s2.

Dur

atio

ntim

e(a

vera

getim

e/da

y)of

cycl

ing

fortran

spor

t

SENo

1.Li

ving

ina

neig

h-bo

rhoo

dw

ithm

ore

stre

etsw

here

the

spee

dlim

itis

≤30

km/h

,mor

epa

rked

cars

that

form

anob

stac

leon

the

road

,mor

etree

s,or

mor

elit

ter

wer

eal

lass

ocia

ted

with

bein

gm

ore

likel

yto

enga

gein

cycl

ing

fortran

s-po

rt.

2.Li

ving

ina

neig

hbor

-ho

odw

ithm

oretraffi

cca

lmin

gfe

atur

es,o

rfe

wer

bicy

cle

lane

s,w

asas

soci

ated

with

bein

gle

sslik

ely

toen

gage

incy

clin

gfo

rtran

spor

t.3.

Livi

ngin

ane

ighb

or-

hood

with

mor

eca

rsth

atfo

rman

obstac

leon

the

road

was

asso

-ci

ated

with

mor

em

in-

utes

ofcy

clin

gfo

rtran

spor

tpe

rw

eek.

4.Li

ving

ina

neig

hbor

-ho

odw

ithm

ore

tree

sw

asas

soci

ated

with

few

erm

inut

esof

cy-

clin

gfo

rtran

spor

tpe

rw

eek.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

15

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

11de

Vrie

set

al.

(201

0)

Net

herlan

ds12

28ch

ildre

nag

ed6–

11ye

arsfrom

10di

sadv

an-

tage

dne

ighb

orho

odsin

Haa

rlem

,Rot

terd

am,

Am

ersfoo

rt,S

chie

dam

,Vl

aard

inge

n,Hen

gelo

.

Obs

erva

tion

base

don

am

odifi

edNei

ghbo

rhoo

dEn

viro

nmen

tW

alka

bilit

ySc

ale

1.Pr

opor

tion

ofre

si-

dent

sto

ente

rprise

s,2.

Prop

ortio

nof

gree

nsp

ace

tore

side

nts,

3.Fr

eque

ncy

ofun

occu

-pi

edho

uses

4.Pr

esen

ceof

gree

nsp

ace

5.Pr

esen

ceof

wat

er6.

Freq

uenc

yof

14ite

ms

onw

alki

ngan

dcy

-cl

ing

infras

truc

ture

7.Gen

eral

impr

ession

ofth

eac

tivity

-frie

ndli-

ness

ofth

ene

ighb

or-

hood

Nei

ghbo

rhoo

din

the

last

7da

ys,

the

num

berof

hour

sof

cycl

ing

tosc

hool

and

cycl

ing

fortran

spor

t

SENo

1.Fo

rcy

clin

gfo

rtran

spor

tatio

n,sig-

nific

antco

rrel

ates

wer

eth

enu

mbe

rof

recr

eatio

nfa

cilit

ies

and

thefreq

uenc

yof

pede

strian

cros

s-in

gs;

2.Cy

clin

gto

scho

olw

asstro

nges

tas

soci

ated

with

the

num

berof

recr

eatio

nfa

cilit

ies,

the

pres

ence

ofgr

een

spac

e,an

dth

efre-

quen

cyof

pede

strian

cros

sing

s,traffi

clig

hts,

and

para

llel

park

ing

spac

esin

the

neig

hbor

hood

;12

Emm

aJ

Ada

-m

set

al.

(201

3)

Uni

ted

King

dom

3516

partic

ipan

tsfrom

Card

iff,K

enilw

orth

and

Sout

ham

pton

Ass

essing

Leve

lsof

Phys

ical

Act

ivity

and

Fitn

ess

Euro

pean

envi

ron-

men

talq

uestio

n-na

ire

1.Cy

clin

gsa

fefrom

traffi

c2.

Safe

tocr

ossro

ads

3.Co

nven

ient

wal

k/cy

cle

rout

es4.

Cycl

ela

nes/

rout

es5.

Variet

yof

wal

k/cy

cle

rout

es6.

Plea

sant

tow

alk/

cycl

e7.

Plac

esto

wal

k/cy

cle

to8.

Ope

nsp

aces

9.Fr

eefrom

litte

r10

.M

any

road

junc

tions

Nei

ghbo

rhoo

dTi

me

dura

tion

(in

hour

san

dm

inut

es)

and

the

tota

ldis-

tanc

eof

cycl

ing

for

tran

spor

tand

cy-

clin

gfo

rre

crea

-tio

n.in

past

wee

k

SE/O

TNo

Cycl

ing

fortran

spor

tw

asas

soci

ated

only

with

stre

etco

nnec

tivity

.

13Hee

sch

etal

.(2

014)

Aus

tral

ia10

,233

adul

tsag

ed40

–65

year

sin

Brisba

neAbb

revi

ated

vers

ion

ofth

eNei

ghbo

rhoo

dEn

viro

nmen

tW

alka

bilit

ySc

ale

1.Aes

thet

ics

2.M

any

traffi

cca

lmin

gde

vice

s3.

Man

ystre

etsar

ehi

lly4.

Man

ycu

l-de-

sacs

5.M

any

4-w

ayin

ters

ec-

tions

Nei

ghbo

rhoo

d1.

Freq

uenc

yof

cycl

ing

forre

-cr

eatio

nor

ex-

erci

sein

the

last

12m

onth

s2.

Tim

edu

ratio

n(h

ours

and

min

utes

)of

cy-

clin

gfo

rtran

s-po

rtin

the

last

wee

k’

SEYe

sLi

ttle

crim

e,fe

wcu

l-de-

sacs

,nea

rby

tran

spor

tan

dre

crea

tiona

ldes

ti-na

tions

wer

eas

soci

ated

with

utili

tycy

clin

g.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

16

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

14Hee

sch

etal

.(2

015)

Aus

tral

ia11

,036

adul

tsag

ed40

–65

year

sfrom

Brisba

neM

apIn

foGIS

data

-ba

sean

dM

apIn

foPr

ofes

sion

al

1.Co

untof

4-w

ayor

mor

ein

ters

ectio

ns2.

Hec

tare

sof

tree

cov-

erag

efrom

aerial

phot

ogra

phy

3.Km

ofoff

-roa

dcy

cle-

way

s4.

Coun

tof

stre

etlig

hts

5.St

anda

rdde

viat

ion

ofhi

lline

ss6.

Ave

rage

reside

ntia

lpa

rcel

size

insq

uare

met

ers

7.Deg

ree

tow

hich

ther

eis

am

ixof

land

use

8.Net

wor

kdi

stan

ceto

CBD

9.Net

wor

kdi

stan

ceto

near

estfe

rry

stop

10.Net

wor

kdi

stan

ceto

near

estrive

r11

.Net

wor

kdi

stan

ceto

near

estsh

op12

.Net

wor

kdi

stan

ceto

near

esttrai

nstat

ion

13.Net

wor

kdi

stan

ceto

near

estco

astli

ne

Nei

ghbo

rhoo

d1.

Tim

edu

ratio

nof

cycl

ing

fortran

s-po

rtin

the

last

wee

k2.

Freq

uenc

yof

cy-

clin

gfo

rre

crea

tion

inla

st12

mon

ths

SE/O

TYe

s1.

Reside

ntsliv

ing

<10

km(n

etw

ork

distan

ce)from

the

Brisba

neCB

Dor

from

the

near

est

ferr

ystop

had

anin

crea

sed

likel

ihoo

dof

cycl

ing

fortran

s-po

rt.

2.Re

side

ntsw

holiv

ed>

5km

from

the

Brisba

neRi

verw

ere

less

likel

yto

cycl

efo

rtran

spor

tth

anw

ere

thos

ew

holiv

ed<

3km

from

the

rive

r.3.

Livi

ngin

ane

ighb

or-

hood

with

am

oder

-at

ely

high

amou

ntof

tree

cove

rage

orof

bike

path

s(Qua

rtile

3)w

asas

soci

ated

with

decr

ease

dlik

elih

ood

ofre

crea

tion-

only

cy-

clin

gco

mpa

red

with

livin

gin

ane

ighb

or-

hood

with

the

high

est

amou

ntsof

thes

eat

-trib

utes

.4.

Reside

ntsliv

ing

<3

kmfrom

the

CBD

wer

em

ore

likel

yto

cycl

efo

rre

crea

tion-

only

com

pare

dw

ithth

ose

livin

g>

10km

from

that

loca

tion.

5.Re

side

ntsliv

ing

<1

kmfrom

the

clo-

sest

shop

sw

ere

less

likel

yto

cycl

efo

rre

-cr

eatio

n-on

lyco

m-

pare

dw

ithth

ose

livin

g>

1km

from

thes

esh

ops.

6.Li

ving

3–5

kmfrom

the

clos

estt

rain

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

17

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts stat

ion

decr

ease

dth

elik

elih

ood

ofcy

clin

gfo

rre

crea

tion-

only

.7.

Thos

eliv

ing

<5

kmfrom

the

coas

thad

anin

crea

sed

likel

ihoo

dof

recr

eatio

n-on

lycy

-cl

ing

com

pare

dw

ithth

ose

livin

g>

10km

from

the

coas

t.15

Fors

yth

and

Oak

es(2

014)

Am

eric

a70

3pa

rtic

ipan

tsag

edab

ove

24from

the

Min

neap

olis-S

t.Pa

ular

eaof

Min

neso

ta

1.Ex

istin

gla

ndus

ean

dbu

sine

ssda

ta-

base

s2.

Aae

rial

phot

oin

-te

rpre

tatio

n3.

Fiel

dwor

k,pr

o-ce

ssed

usin

gth

eArc

GIS

,4.S

urve

y

1.Po

pula

tion/

land

area

(net

wor

kof

200

m,

400

m,s

trai

ghtl

ine

of20

0m

)2.

Popu

latio

n/de

velo

ped

land

area

(400

mne

t-w

ork,

200

mstra

ight

line)

3.Po

pula

tion

and

em-

ploy

men

t/la

ndar

ea(4

00m

netw

ork,

200

mstra

ight

line)

4.Hou

sing

units

/uni

tla

ndar

ea(2

00m

net-

wor

k,40

0m

netw

ork)

5.Pa

rcel

area

:Ind

ustria

l(%

)(4

00m

netw

ork)

6.Pa

rcel

area

:Par

ks/

Rec

(%)(

1,60

0m

stra

ight

line)

7.Pa

rcel

area

:Ind

ustria

lan

dau

to-o

rien

ted

uses

(%)(4

00m

net-

wor

k)8.

Stre

etse

gmen

tsw

ithvi

sibl

elit

ter,

graffi

ti,or

dum

pste

rs(%

)(2

00m

netw

ork)

9.Re

side

ntia

lPo

pula

tion/

reside

n-tia

lpar

cela

rea

10.Lo

tcov

erag

e11

.M

edia

nce

nsus

bloc

kar

ea12

.In

ters

ectio

nspe

run

itar

ea

Indi

vidu

alre

spon

-de

nt1.

Num

bero

fday

san

ddu

ratio

ntim

eof

cycl

ing

atle

ast10

min

during

the

last

7da

ys.

2.Tr

ipsbe

twee

ntw

opl

aces

with

distan

cesre

-co

rded

inm

iles

orbl

ocks

.3.

Stat

eth

em

ost

rece

nttim

eth

eyha

dridd

enin

neig

hbor

-ho

odif

peop

leha

dcy

cled

inth

eirne

ighb

or-

hood

inth

epa

sttw

oye

ars

SE/O

TNo

1.M

ore

cycl

ing

oc-

curr

edw

here

peop

leliv

edin

dustrial

,au

to-o

rien

ted

land

uses

,and

inar

eas

with

visibl

elit

ter

and

graffi

ti.2.

Ina

buffe

rof

1600

mth

epa

rkar

eais

asso

-ci

ated

with

tota

lmile

scy

cled

3.Lo

wer

dens

ities

,lar

ger

bloc

ks,f

ewer

side

-w

alks

,and

few

erbu

sstop

sar

eas

soci

ated

with

the

freq

uenc

yof

cycl

ing,

thos

eliv

ing

inlo

wer

dens

ityar

eas

cycl

ing

mor

e.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

18

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

13.M

edia

npe

rim

eter

ofbl

ock

14.Pr

opor

tion

ofdi

ssim

-ila

rla

ndus

esam

ong

grid

15.Si

dew

alk

leng

th/r

oad

leng

th16

.Tr

ansitstop

dens

ity17

.Acc

essto

serv

ices

18.Pl

aces

forw

alki

ngan

dcy

clin

g19

.Nei

ghbo

rhoo

dsu

r-ro

undi

ngs/

attrac

tive-

ness

20.Sa

fety

from

traffi

c16

Bueh

leran

dPu

cher

(201

1)

Am

eric

a90

ofth

e10

0la

rges

tU.S

.ci

tiesas

dete

rmin

edby

popu

latio

nes

timat

esof

the

2008

Am

eric

anCo

mm

unity

Surv

ey

The

Leag

ueof

Am

eric

anBi

cycl

ists

and

the

Alli

ance

for

Biki

ngan

dW

alki

ng

1.M

ilesof

bike

lane

sin

city

per10

0,00

0po

-pu

latio

n2.

Mile

sof

bike

and

shar

ed-u

sepa

thsin

city

per10

0,00

0po

-pu

latio

n3.

Regi

onal

inde

xco

m-

bini

ng22

variab

les

mea

suring

reside

ntia

lde

nsity

,lan

dus

em

ix,

stre

ngth

ofdo

wn-

tow

ns,a

ndco

nnec

-tiv

ityof

stre

etne

t-w

ork,

city

1.Sh

are

ofcy

-cl

ing

forco

m-

mut

ing

inea

chci

ty;a

nd2.

The

num

bers

ofbi

keco

mm

u-te

rspe

r10

,000

popu

latio

n

SE/W

E/LS

No

Citie

sw

itha

grea

ter

supp

lyof

bike

path

sand

lane

sha

vesign

ifica

ntly

high

erbi

keco

mm

ute

rate

s

17M

oran

etal

.(2

015)

Isra

el57

3ch

ildre

nag

ed10

–12

year

s(fi

fthor

sixt

hgr

ade)

from

Rish

onLe

Zion

GIS

Ina

400

mbu

ffer,

1.Re

side

ntia

lden

sity

(num

berof

hous

e-ho

ldspe

rsq

uare

kilo

-m

eter

),2.

Built

cove

rage

(ove

rall

built

area

per

built

lots

area

)3.

Stre

etco

nnec

tivity

(num

berof

inte

rsec

-tio

nspe

rsq

uare

kilo

-m

eter

)4.

LUM

mea

sure

s:a.

Rout

edi

stan

ceto

near

ests

tore

b.Ro

ute

distan

ceto

near

estp

ark

Nei

ghbo

rhoo

d1.

Cycl

ing

tone

ighb

orho

odde

stin

atio

nat

leas

tone

time

aw

eek

2.Cy

clin

gfo

rle

i-su

reat

leas

ton

etim

ea

wee

k

SEYe

sLi

ving

inan

area

with

low

reside

ntia

lden

sity

isas

soci

ated

with

the

incr

ease

dod

dsof

both

cycl

ing

tone

ighb

orho

odan

dle

isur

ecy

clin

g

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

19

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

c.Ro

ute

distan

ceto

scho

ol18

Salli

set

al.

(201

3)

Am

eric

a17

80ad

ults

aged

20–6

5re

crui

ted

from

the

Seat

tle,

Was

hing

ton

and

Balti

mor

e,M

aryl

and

re-

gion

s

1.GIS

2.Nei

ghbo

rhoo

dEn

viro

nmen

tW

alka

bilit

ySc

ale

1.Nei

ghbo

rhoo

dstre

ets

are

hilly

,wal

king

isdi

fficu

lt2.

Bike

/ped

estria

ntrai

lsar

eea

syto

getto

3.Sa

feto

ride

bike

inne

ighb

orho

od4.

Reside

ntia

lden

sity

5.La

ndus

em

ix-d

iver

-sity

6.La

ndus

em

ix-a

cces

s7.

Stre

etco

nnec

tivity

8.W

alki

ng/c

yclin

gfa

cil-

ities

9.Nei

ghbo

rhoo

dae

s-th

etic

s10

.Pe

destrian

/tra

ffic

safe

ty11

.Net

reside

ntia

lden

sity

(ln-

tran

sfor

med

)12

.In

ters

ectio

nde

nsity

13.Re

tail

floor

area

ratio

14.La

ndM

ixed

use

15.W

alka

bilit

yin

dex

Nei

ghbo

rhoo

dBi

cycl

ing

freq

uenc

ySE

No

Hig

herridi

ngfreq

uenc

yis

asso

ciat

edw

ith1.

Hav

ing

bike

/ped

es-

tria

ntrai

lsea

syto

get

to,

2.Gre

ater

safe

tyfo

rridi

ngin

the

neig

h-bo

rhoo

d3.

Gre

ater

land

use

mix

-ac

cess

.4.

Nei

ghbo

rhoo

dw

alk-

abili

tym

easu

res

with

in1

kmw

ere

not

cons

iste

ntly

rela

ted

tobi

cycl

ing.

19Nie

lsen

etal

.(2

013)

Den

mar

k91

28cy

clists

aged

10–8

5ye

arsfrom

Dan

ish

Nat

iona

lTra

vels

urve

y

1.Dan

ish

build

ing

regi

ster

2.Co

rine

land

cove

rda

tase

ts3.

Nav

teq

road

netw

ork

4.Dig

itale

leva

tion

mod

el

ina

500

man

d15

00m

buffe

r1.

Den

sity

ofpo

pula

tion,

jobs

and

reta

iljo

bs.

2.Th

epe

rcep

tion

ofus

eof

land

forur

ban

area

s;fo

rest

and

natu

rear

eas;

land

use

variat

ion

base

don

44la

ndus

ecl

asse

s,th

ebl

end

ofpo

pula

tion,

jobs

and

reta

il.3.

Num

beran

dde

nsity

ofth

ree-

way

inte

rsec

-tio

ns;t

heco

mpo

sitio

nof

the

road

netw

ork

(tra

ffic

road

s,di

stri-

buto

rro

ads,

and

loca

lro

ads)

;and

the

pro-

portio

nof

build

ings

inth

ear

eath

atw

ere

Nei

ghbo

rhoo

dDista

nce

ofcy

clin

gtrip

spe

rda

ySE

/OT/

STYe

s1.

Flat

terr

ain,

shor

t-di

stan

ceto

reta

ilco

ncen

trat

ions

,as

wel

laspo

pula

tion

dens

ityan

dne

twor

kco

nnec

tivity

with

ina

1.5

km‘p

erso

nal’

neig

hbor

hood

,all

cont

ribu

teto

anin

-cr

ease

dlik

elih

ood

ofcy

clin

g.2.

Publ

ictran

spor

tatio

nac

cess

and

leve

l-of-

serv

ice,

mea

sure

das

the

loca

tion

ofa

trai

nstat

ion

with

in10

00m

and

the

num

berof

daily

busan

dtrai

nde

partur

esw

ithin

500

mof

the

hom

e,

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

20

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

built

befo

re19

50is

used

asan

indi

cato

rof

para

met

eriz

atio

nur

ban

design

.4.

Acc

essto

the

near

est

prim

ary

scho

olan

dgr

ocer

ysh

ops,

acce

ssto

job

and

reta

ilco

n-ce

ntra

tions

ofsize

san

dsc

ales

.5.

Publ

ictran

spor

tatio

nstop

san

dtrai

nan

dbu

sdep

artu

resb

ystop

6.On-

site

land

use

in-

tens

itydi

stan

ceto

am

otor

way

ram

por

larg

ero

ad.

both

redu

ceth

elik

e-lih

ood

ofcy

clin

g.3.

Reta

iljo

bspe

rre

si-

dent

with

ina

very

conv

enie

nt50

0m

wal

king

rang

ein

di-

cate

sm

ain

stre

etsan

dce

ntra

lare

alo

catio

ns,

and

toge

ther

with

the

netw

ork

conn

ectiv

ityw

ithin

the

500

mra

nge,

isne

gativ

ely

corr

elat

edto

the

prob

abili

tyof

cycl

ing.

4.Lo

ngdi

stan

cesto

larg

ere

tail

conc

entra-

tions

,hig

hpo

pula

tion

dens

ity,a

ndhi

ghne

t-w

ork

conn

ectiv

ityw

ithin

wal

king

rang

ear

ere

late

dto

shor

tda

ilycy

clin

gdi

s-ta

nces

.20

Van

Hol

leet

al.(

2014

)

Belg

ium

59ad

ults

aged

45–6

4ye

arsre

side

din

sem

i-ur

ban

(300

–600

inha

bi-

tant

s/km

2)or

urba

n(6

00in

habi

tant

s/km

2)ne

igh-

borh

oods

from

Brus

sels

Pano

ram

icco

lor

phot

ogra

phsof

sem

i-urb

anan

dur

ban

stre

etsc

apes

1.Ope

nnes

sof

view

2.Cy

cle

path

wid

th3.

Cycl

epa

thev

enne

ss4.

Pres

ence

ofdr

ivew

ays

cros

sing

cycl

epa

th5.

Safe

tycr

ossing

the

stre

et6.

Pres

ence

ofhi

stor

icel

emen

ts7.

Pres

ence

ofne

wel

e-m

ents

8.Ve

geta

tion

9.Upk

eep

ofth

ecy

cle

path

10.Upk

eep

ofth

eside

-w

alk

11.Gen

eral

upke

ep12

.Nat

ural

surv

eilla

nce

13.La

ndus

e14

.Num

berof

traffi

cla

nes

15.Se

para

tion

cycl

epa

than

dside

wal

k

Indi

vidu

alre

spon

-de

ntTi

me

dura

tion

(min

s)of

cycl

ing

fortran

spor

tatio

npe

rw

eek

SEYe

sPr

esen

ceof

vege

tatio

nw

asth

em

ostim

portan

ten

viro

nmen

talc

hara

c-te

ristic

toin

vite

peop

lefo

ren

gagi

ngin

tran

s-po

rtat

ion

cycl

ing.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

21

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

21Su

net

al.

(201

7)

Scot

land

1368

4St

rava

cycl

ists

inGla

sgow

Clyd

eVa

lley

Plan

ning

Are

a

1.Ope

nStree

tMap

2.St

rava

Met

roda

tase

t3.

Scot

land

's20

11ce

nsus

data

4.DATA

.GOV.

UK

5.Eu

rope

anEn

viro

nmen

tAge

ncy

6.Th

eUK

Dep

artm

entfo

rTr

ansp

ort

1.Dista

nce

toci

tyce

nter

2.Dista

nce

toth

ene

ares

tbu

sstop

3.Ro

adle

ngth

4.Co

nnec

tivity

ofm

ajor

road

5.Co

nnec

tivity

ofm

inor

road

6.Ro

adcl

ass

7.La

ndus

em

ix8.

Dom

inan

tlan

dus

ety

pe9.

Cont

igui

tyto

gree

nsp

ace

10.Vo

lum

eof

mot

orve

-hi

cles

11.Tr

affic

acci

dent

den-

sity

12.Po

pula

tion

dens

ity13

.Em

ploy

men

tden

sity

Indi

vidu

alre

spon

-de

ntRe

crea

tiona

lCy

clin

gRa

te(R

CR),

the

rate

ofre

crea

tiona

ltrips

during

a1-

htim

eslot

SENo

1.Ro

adle

ngth

hasa

sign

ifica

ntan

dne

-ga

tive

asso

ciat

ion

with

RCR

2.Bo

thco

nnec

tivity

ofm

ajor

road

and

con-

nect

ivity

ofm

inor

road

are

positiv

ely

and

sign

ifica

ntly

asso

-ci

ated

with

RCR.

3.Oft

hedo

min

antla

ndus

ety

pe,“

Reside

ntia

l”ha

sa

sign

ifica

ntan

dpo

sitiv

eas

soci

atio

nw

ithRC

R.

22Za

habi

etal

.(2

016)

Cana

daOrigi

n–de

stin

atio

n(O

–D)

surv

eyda

tafrom

Gre

ater

Mon

trea

lin

the

year

1998

(n=

21,1

88),

2003

(n=

20,1

70)an

d20

08(n

=19

,508

)

Stat

istic

sCa

nada

for

cens

usye

ars,

Mon

trea

land

Vélo

Qué

bec,

Des

ktop

Map

ping

Tech

nolo

gies

Inc.,

GIS

in50

0m

grid

system

,1.

Popu

latio

nde

nsity

;2.

Empl

oym

entd

ensity

;3.

Cycl

ing

netw

ork

den-

sity

;4.

Publ

ictran

sita

cces

si-

bilit

y;5.

Land

-use

mix

6.In

ters

ectio

n(n

ode)

dens

ity7.

Stre

etde

nsity

8.Li

nkto

node

ratio

9.Dista

nce

tocy

clin

gin

fras

truc

ture

Nei

ghbo

rhoo

dSh

are

ofcy

clin

gch

oice

SEYe

s1.

A10

%de

crea

sein

distan

ceto

near

est

cycl

ing

faci

litie

s(m

easu

red

inkm

)w

ould

onav

erag

ere

sult

ina

3.7%

in-

crea

sein

the

prob

-ab

ility

ofch

oosing

tocy

cle.

2.In

ters

ectio

nde

nsity

and

the

link

tono

dera

tioof

reside

ntia

llo-

catio

nw

ere

foun

dto

bepo

sitiv

ely

asso

-ci

ated

with

the

choi

ceof

cycl

ing

tow

ork

Colo

mbi

aNei

ghbo

rhoo

dSE

/OT

No

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

22

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

23Ce

rver

oet

al.

(200

9)

830

adul

tsag

edov

er18

year

sw

hore

ported

that

they

know

how

toride

abi

kein

Bogo

ta

Cada

ster

Dep

artm

entof

the

City

ofBo

gota

usin

gGeo

grap

hic

Info

rmat

ion

System

s(G

IS)to

ols.

1.Dw

ellin

gun

itspe

rhe

ctar

e;2.

%of

land

area

occu

-pi

edby

build

ings

;3.

Ave

rage

build

ing

floor

heig

ht;

4.Pl

otra

tio(b

uild

ing

m2

=la

ndm

2)5.

Entrop

yin

dex

ofla

nd-

use

mix

(0–1

scal

e);

6.Pr

opor

tion

ofbu

ild-

ings

vertic

ally

mix

ed;

7.Pr

opor

tion

ofto

tal

floor

spac

ein

build

-in

gsw

ith2

+us

es;

8.Pu

blic

park

area

as%

ofto

tall

and

area

;9.

Ave

rage

park

size

(hec

tare

s);

10.%

ofro

adlin

ksw

ithm

edia

nstrips

;11

.Tr

affic

light

dens

ity(tra

ffic

light

s/stre

etle

ngth

);12

.Tr

eede

nsity

(tre

es/

stre

etle

ngth

);13

.Ave

rage

lots

ize

(m2)

;14

.Qua

drila

tera

llot

sas

%of

tota

l;15

.Pe

rcen

tofb

lock

sw

ithco

ntai

ned

hous

ing

and

acce

ssco

ntro

l;16

.St

reet

dens

ity(s

tree

tar

ea/l

and

area

);17

.Pr

opor

tion

ofin

ter-

sect

ions

with

:1po

int

(cul

desa

c),3

poin

ts,

4po

ints,5

+po

ints;

18.Bi

ke-la

nede

nsity

19.Ro

ute

dire

ctne

ss20

.Co

nnec

tivity

inde

x(int

erse

ctio

nno

des/

stre

etlin

ks);

21.Num

berof

brid

ges;

22.Ci

clov

iatw

o-w

ayle

ngth

(lin

ealm

eter

s);

Tim

edu

ratio

nof

cycl

ing

fortran

s-po

rtpe

rda

y

1.Hig

hstre

etde

nsiti

es(ie.

road

km/l

and

area

km>

0.20

)w

illin

crea

sene

arly

twic

eas

likel

yto

cycl

efo

rut

ilita

rian

purp

oses

30m

inor

mor

epe

rw

eekd

ay.

2.St

eep

topo

grap

hyal

sode

ters

cycl

ing.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

23

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

23.Num

berof

pede

strian

brid

ges;

24.Pe

destrian

acci

dent

spe

rye

ar;

25.Ave

rage

auto

mob

ilesp

eeds

onm

ain

stre

ets;

26.Dea

ths(a

llty

pes)

intraffi

cac

cide

ntspe

rye

ar;

27.Num

berof

repo

rted

crim

espe

rye

ar;

28.Num

berof

:pub

licsc

hool

s;ho

spita

ls;

publ

iclib

raries

;sho

p-pi

ngce

nter

s(>

500m

2);c

hurc

hes;

bank

s;29

.Num

berof

Tran

sM

ileni

o(B

RT)sta-

tions

;30

.Sh

orte

stne

twor

kdi

s-ta

nce

tocl

oses

tTra

nsM

ileni

ostat

ion;

31.Num

berof

feed

erTr

ansM

ileni

ostat

ions

24Ch

ristia

nsen

etal

.(20

16)

Ade

laid

e,Aus

tral

ia(A

US)

;Ghe

nt,B

elgi

um(B

EL);

Curitib

a,Br

azil

(BRA

);Bo

gota

,Col

ombi

a(C

OL)

;Olo

mou

c,Cz

ech

Repu

blic

(CZ)

;Aar

hus,

Den

mar

k(D

EN);

Cuer

nava

ca,

Mex

ico

(MEX

),Ch

ristch

urch

,Nor

thSh

ore,

Wai

take

re,a

ndW

ellin

gton

,New

Zeal

and

(NZ)

,Sto

ke-o

n-Tr

ent,

Uni

ted

King

dom

(UK)

and

Balti

mor

ean

dSe

attle

,Uni

ted

Stat

es(U

S)

12,1

81ad

ults

aged

18–6

6ye

arsfrom

14ci

ties

GIS

1.Net

reside

ntia

lden

sity

(num

berof

dwel

lings

perkm

2of

buffe

rar

easde

vote

dto

resi-

dent

ialu

se),

2.La

nd-u

sem

ix(e

ntro

pysc

ore

ofth

ree

land

-us

es:r

esid

entia

l,re

tail

and

civi

c)3.

Stre

etco

nnec

tivity

(num

berof

inte

rsec

-tio

nsw

ithth

ree

orm

orein

ters

ectin

gro

adse

gmen

tspe

rkm

2)4.

Park

s(n

umbe

rof

park

sin

ters

ectin

gpa

rtic

ipan

tbuff

ers)

.

City

1.Th

efreq

uenc

yof

cycl

ing

for

tran

spor

t.2.

The

num

berof

days

ofcy

clin

gdu

ring

the

last

seve

nda

ys(a

tle

ast10

min

)

SEYe

s1.

Net

reside

ntia

lden

-sity

,int

erse

ctio

nde

nsity

,and

land

use

mix

wer

epo

si-

tivel

yas

soci

ated

with

the

odds

ofen

-ga

ging

inan

ybo

uts

(>10

min

)of

cy-

clin

gfo

rtran

spor

tin

the

last

wee

k;2.

Onl

yin

ters

ectio

nde

n-sity

and

land

use

mix

rem

aine

dsign

ifica

ntly

asso

ciat

edin

the

mul

ti-en

viro

nmen

t-va

riab

lem

odel

s.3.

Land

use

mix

wer

elin

early

positiv

ely

as-

soci

ated

with

days

per

wee

kof

cycl

ing

for

tran

spor

t.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

24

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

4.Th

enu

mbe

rof

park

s(1

park

/km

2;50

0m

buffe

r)w

erepo

sitiv

ely

asso

ciat

edw

ithth

eod

dsof

any

cycl

ing

for

tran

spor

tin

Ghe

nt(B

EL),

Aar

hus(D

EN)

and

Seat

tle(U

SA)o

nly

25Sn

izek

etal

.(2

013)

Den

mar

k39

8cy

clin

gro

utes

inCo

penh

agen

and

Fred

erik

sber

g

GIS

1.Ro

adsw

ithcy

cle

fa-

cilit

ies.

2.Dista

nce

tone

ares

tcy

cle

rack

.3.

Num

berof

cycl

era

cks

with

in10

0m

.4.

Dista

nce

toth

ene

ares

ttraffi

clig

hts.

5.St

reet

type

s6.

Dista

nce

toth

ecl

oses

tgr

oup

ofbu

sstop

s.7.

Dista

nce

tone

ares

tin

ters

ectio

n.8.

Dista

nce

toto

wn

hall.

9.Num

berof

com

pani

esw

ithin

adi

stan

ceof

100

mto

the

expe

ri-

ence

poin

t10

.Num

berof

reta

ilun

itsw

ithin

adi

stan

ceof

100

mto

the

expe

ri-

ence

poin

t.11

.Dista

nce

tocl

oses

tw

ater

orgr

een

area

.12

.Pe

rcen

tage

ofgr

een

ofro

ute

segm

ents.

13.Dev

iatio

nsfrom

the

dire

ctlin

e.14

.Th

ean

gle

betw

een

the

curr

entse

gmen

tan

dth

elin

efrom

the

expe

rien

cepo

intto

-w

ards

tow

nha

ll.

Indi

vidu

alre

spon

-de

ntPo

sitiv

ean

dne

ga-

tive

cycl

ing

poin

tOT

No

1.Po

sitiv

ecy

clin

gex

-pe

rien

cescl

early

re-

late

dto

the

pre-

senc

eof

en-rou

tecy

clin

gfa

cilit

ies,

and

attrac

tive

natu

reen

viro

n-m

ents

with

ina

shor

tdi

stan

ceof

larg

ew

ater

bodi

esor

gree

ned

gesal

ong

the

rout

e.2.

Neg

ativ

eex

perien

ces

are

rela

ted

tobu

sstop

s,hi

ghtraffi

cde

nsiti

esal

ong

the

rout

e,as

wel

lassig-

nale

dan

dno

n-sig-

nale

din

ters

ectio

ns

26W

inte

rset

al.

(201

0)

Cana

da14

02cu

rren

tan

dpo

ten-

tialc

yclis

tsag

edab

ove

18from

Met

roVa

ncou

ver

1.Tr

ansL

ink

2.Th

ere

gion

altran

spor

tatio

nau

thor

ity,

3.Bi

cycl

eco

ordi

-na

tors

from

the

1.Ve

hicl

es;

2.La

nem

arki

ngs;

3.In

ters

ectio

ns;

4.Dista

nces

,hill

san

dco

nnec

tions

;

Nei

ghbo

rhoo

dCy

clin

gin

tent

OT

No

1.Th

estro

nges

tm

oti-

vato

rsw

ere

rout

esth

atw

ere

away

from

traffi

c,ae

sthe

-tic

ally

plea

sing

,and

easy

tocy

cle.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

25

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

partic

ipat

ing

mun

icip

aliti

es4.

Mem

bers

ofcy

-cl

ing

advo

cacy

grou

ps

5.Ro

adsu

rfac

esan

dm

aint

enan

ce;

6.Aes

thet

icsan

dac

cess

;7.

Coor

dina

tion

with

tran

sit;

8.Sa

fety

2.Th

estro

nges

tde

ter-

rent

sw

ere

unsa

fesu

r-fa

cesan

din

tera

ctio

nsw

ithm

otor

vehi

cles

.

27Va

nDyc

ket

al.(

2010

)

Belg

ium

1166

partic

ipan

tsag

ed20

–65

year

sfrom

24ne

ighb

orho

odsin

Ghe

nt

Serv

ice

for

Envi

ronm

enta

lPl

anni

ngin

Ghe

nt,

GIS

1.W

alka

bilit

yin

dex

cond

ucte

dby

reside

n-tia

lden

sity

,2.

Inte

rsec

tion

dens

ity3.

Land

use

mix

Nei

ghbo

rhoo

dNum

berof

days

and

time

dura

tion

(hou

rsan

dm

inut

espe

rda

y)of

cycl

ing

fortran

spor

tin

last

7da

ys

SENo

Livi

ngin

ahi

gh-w

alk-

able

neig

hbor

hood

wa-

sass

ocia

ted

with

sign

ifi-

cant

lym

ore

cycl

ing

for

tran

spor

t.

28Pi

atko

wsk

ian

dM

arsh

all

(201

5)

Am

eric

a20

30pa

ticip

ants

inDen

ver

GIS

,sur

vey

ina

20fe

etbu

ffer,

1.One

-way

trip

distan

ce(s

elf-r

epor

t)2.

Origi

nlin

k-to

-nod

era

tio3.

Origi

nin

ters

ectio

nde

nsity

4.Des

tinat

ion

link-

to-

node

ratio

5.Des

tinat

ion

inte

rsec

-tio

nde

nsity

6.Sa

me

Origi

nan

dDes

tinat

ion.

7.To

ofe

wbi

kela

nes;

8.La

ckof

conn

ectio

nsbe

twee

nbi

kela

nes

and

path

s;9.

Wor

ries

abou

tro

adsa

fety

(with

rega

rdto

traffi

c);

10.St

reet

cond

ition

s(p

otho

les,

crac

ks,b

ike

lane

s,et

c);

11.Not

enou

ghlig

hton

existin

gbi

cycl

efa

cil-

ities

atni

ght.

Nei

ghbo

rhoo

dCy

clin

gin

tent

SENo

one-

way

trip

distan

ceis

sign

ifica

ntly

nega

tivel

yas

soci

ated

with

both

the

deci

sion

tobe

gin

com

-m

ute-

cycl

ing

asw

ella

sth

ede

cision

toco

mm

ute

bybi

kew

ithgr

eate

rfreq

uenc

y.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

26

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

29Co

le-H

unte

ret

al.(

2015

)

Spai

n76

9ad

ults

aged

abov

e18

year

sin

Barc

elon

aGIS

ina

400

mbu

fferof

hom

ean

dW

ork/

stud

y,1.

Bicy

cle

rack

s2.

Publ

icbi

cycl

estat

ions

3.Pu

blic

tran

spor

tsta

-tio

ns4.

Bicy

cle

lane

(%)

5.Gre

enne

ss,N

DVI

6.El

evat

ion

Indi

vidu

alre

spon

-de

nt1.

Cycl

ing

inte

nt2.

Cycl

ing

fre-

quen

cy

SENo

1.Si

gnifi

cant

positiv

ede

term

inan

tsof

prop

ensity

forbi

-cy

cle

com

mut

ing

wer

eth

equ

antit

yof

publ

icbi

cycl

esta-

tions

with

inth

eho

me

area

,and

amou

ntof

gree

n-ne

ssw

ithin

the

wor

k/stud

yar

ea;

2.Th

equ

antit

yof

publ

ictran

s-po

rtstat

ions

with

inth

eho

me

area

and

the

mea

nel

eva-

tion

ofth

ew

ork/

stud

yar

eaw

ere

sign

ifica

ntne

gativ

ede

term

inan

tsof

the

sam

epr

open

-sity

.30

Mer

tens

etal

.(2

016)

Belg

ium

,the

Net

herlan

ds,

Hun

gary

,Fra

nce,

UK

4612

adul

tsfrom

five

Euro

pean

regi

ons,

Ghe

ntre

gion

(Bel

gium

),Ra

ndstad

regi

on(the

Net

herlan

ds),

Buda

-pes

t(H

unga

ry),

Parisre

gion

(Fra

nce)

and

Gre

ater

Lond

on(U

K)

Ass

essing

Leve

lsof

Phys

ical

Act

ivity

envi

ronm

enta

lqu

estio

nnai

re

Ina

five-

poin

tLik

erts

cale

:1.

Ther

ear

esp

ecia

lla

nes,

rout

esor

path

sfo

rcy

clin

gin

my

neig

hbor

hood

;2.

Ther

eis

heav

ytraffi

cin

my

neig

hbor

hood

rela

ted

tocy

clin

g;3.

Thecy

cles

path

sin

my

neig

hbor

hood

are

wel

lm

aint

aine

d;4.

My

neig

hbor

hood

isa

plea

sant

area

for

wal

king

orcy

clin

g;5.

My

neig

hbor

hood

isge

nera

llyfree

from

litte

r,w

aste

orgr

affiti;

6.Th

eai

rin

my

neig

h-bo

rhoo

dis

pollu

ted;

7.Th

esp

eed

oftraffi

cin

my

neig

hbor

hood

isus

ually

low

;8.

The

leve

lofc

rim

ein

my

neig

hbor

hood

ishi

ghan

d9.

Ihav

ea

choi

ceof

dif-

fere

ntro

utes

for

Nei

ghbo

rhoo

dnu

mbe

rofd

aysa

nddu

ratio

n(a

vera

getim

epe

rda

y)of

cycl

ing

fortran

s-po

rtin

last

7da

ys

SE/O

TNo

1.Lo

wpe

rcei

ved

traffi

csp

eed,

high

perc

eive

dch

oice

betw

een

diffe

rent

rout

esto

wal

k/cy

cle

and

high

perc

eive

dai

rpo

llutio

nin

the

neig

hbou

rhoo

dw

ere

positiv

ely

as-

soci

ated

with

the

odds

ofen

gagi

ngin

cycl

ing

fortran

s-po

rt.

2.Pe

rcei

ved

plea

sant

-ne

ssof

the

envi

ron-

men

tin

rela

tion

tow

alki

ngor

cycl

ing

asw

ella

spe

rcei

ved

less

airpo

llutio

nw

ere

ne-

gativ

ely

asso

ciat

edw

ithm

inut

espe

rw

eek

cycl

ing

for

tran

spor

tam

ong

thos

ew

hoin

dica

ted

toha

vecy

cled

inth

ela

stse

ven

days

.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

27

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

wal

king

orcy

clin

gin

my

neig

hbor

hood

.31

Del

fien

Van

Dyc

ket

al.

(201

2)

Am

eric

a,Aus

tral

ia,

Belg

ium

6014

adul

tsag

ed20

–65

year

sfrom

the

USA

(Bal

timor

ean

dSe

attle

),Aus

tral

ia(A

dela

ide)

and

Belg

ium

(Ghe

nt)

the

Dut

chan

dEn

glish

vers

ions

ofth

epr

evio

usly

vali-

date

dNei

ghbo

rhoo

dEn

viro

nmen

tal

Wal

kabi

lity

Scal

e(N

EWS)

1.Re

side

ntia

lden

sity

2.La

ndus

em

ixdi

vers

ity–

prox

imity

ofde

sti-

natio

ns3.

Land

usem

ixdi

vers

ity-#

destin

atio

nsw

ithin

20m

inw

alk

4.La

ndus

em

ixac

cess

5.Not

man

ycu

l-de-

sacs

6.Pa

rkin

gdi

fficu

ltne

arlo

cals

hopp

ing

area

7.Not

man

yba

rrie

rsin

neig

hbor

hood

8.St

reet

conn

ectiv

ity9.

Prox

imity

totran

sit

stop

10.W

alki

ngan

dcy

clin

gfa

cilit

ies

Nei

ghbo

rhoo

dNum

berof

days

and

dura

tion

(ave

rage

time

per

day)

ofcy

clin

gfo

rtran

spor

tin

last

7da

ys

SE/O

TNo

Prox

imity

tode

stin

a-tio

ns,g

ood

wal

king

and

cycl

ing

faci

litie

s,pe

r-ce

ivin

gdi

fficu

lties

inpa

rkin

gne

arlo

cal

shop

ping

area

s,an

dpe

rcei

ved

aesthe

tics

wer

ein

clud

edin

a‘cy-

clab

ility

’ind

ex.T

his

inde

xw

aslin

early

posi-

tivel

yre

late

dto

tran

s-po

rt-rel

ated

cycl

ing

and

noge

nder

-orco

untry-

diffe

renc

esw

ere

ob-

serv

ed.

32Ke

rret

al.

(201

6)

Aus

tral

ia,B

elgi

um,B

razi

l,Ch

ina,

Colo

mbi

a,Cz

ech

Repu

blic

,Den

mar

k,M

exic

o,New

Zeal

and,

Spai

n,th

eUni

ted

King

dom

,the

Uni

ted

Stat

es

13,7

45ad

ults

aged

18–6

5ye

arsfrom

17ci

tiesin

12co

untrie

s,in

clud

ing

Aus

tral

ia(A

US)

:Ade

laid

e;Be

lgiu

m(B

EL):

Ghe

nt;

Braz

il(B

R):C

uriti

ba;

Chin

a(C

N):

Hon

gKo

ng;

Colo

mbi

a(C

OL)

:Bog

ota;

Czec

hRe

publ

ic(C

Z):

Olo

mou

can

dHra

dec

Kral

ove;

Den

mar

k(D

EN):

Aar

hus;

Mex

ico

(MEX

):Cu

erna

vaca

;New

Zeal

and

(NZ)

:Nor

thSh

ore,

Wai

take

re,W

ellin

gton

,an

dCh

ristch

urch

;Spa

in(S

P):P

ampl

ona;

the

Uni

ted

King

dom

(UK)

:St

oke-

on-T

rent

;and

the

Uni

ted

Stat

es(U

S):

Seat

tle,W

ashi

ngto

n,an

dBa

ltim

ore,

Mar

ylan

d.

Nei

ghbo

rhoo

dEn

viro

nmen

tW

alka

bilit

ySc

ale–

Abb

revi

ated

(NEW

S–A):

IPEN

Subs

cale

san

dIte

ms

1.How

com

mon

are

de-

tach

edsing

le-fa

mily

reside

nces

/To

wnh

ouse

sor

row

sof

1–3-

stor

yho

uses

/Apa

rtm

ents

orco

ndos

with

1–3

stor

ies/

Apa

rtm

ents

orco

ndos

with

4–6

stor

ies/

Apa

rtm

ents

orco

ndos

with

7–12

stor

ies/

Apa

rtm

ents

orco

ndos

with

>12

stor

ies/

Apa

rtm

ents

orco

ndos

with

>20

stor

ies

2.St

ores

are

with

inea

syw

alki

ngdi

stan

ceof

my

hom

e.3.

Ther

ear

em

any

plac

esto

gow

ithin

easy

wal

king

distan

ceof

my

hom

e.4.

Itis

easy

tow

alk

toa

tran

sitstop

(bus

,trai

n)from

my

hom

e.5.

The

distan

cebe

twee

nin

ters

ectio

nsin

my

City

1.Sh

are

ofcy

-cl

ing

fortran

s-po

rtpo

pula

-tio

n,2.

Dur

atio

ntim

eof

cycl

ing

for

tran

spor

tin

last

7da

ys

SEYe

s1.

Asign

ifica

ntno

n-lin

earas

soci

atio

nbe

twee

npe

rcei

ved

reside

ntia

lden

sity

and

any

cycl

ing

for

tran

spor

tth

atw

asco

nsiste

ntly

nega

-tiv

ein

slop

e;2.

Any

cycl

ing

fortran

s-po

rtw

assign

ifica

ntly

rela

ted

tope

rcei

ved

land

use

mix

–acc

ess,

stre

etco

nnec

tivity

,in

fras

truc

ture

,aes

-th

etic

s,sa

fety

,and

perc

eive

ddi

stan

ceto

destin

atio

ns.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

28

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

neig

hbor

hood

isus

ually

shor

t6.

Ther

ear

em

any

alte

r-na

tive

rout

esfo

rge

t-tin

gfrom

plac

eto

plac

ein

my

neig

hbor

-ho

od7.

Ther

ear

eside

wal

kson

mos

tof

the

stre

ets

inm

yne

ighb

orho

od.

8.M

yne

ighb

orho

odstre

etsar

ew

elll

itat

nigh

t.9.

Wal

kers

and

bike

rson

the

stre

etsin

my

neig

hbor

hood

can

beea

sily

seen

bype

ople

inth

eirho

mes

.10

.Th

ere

are

cros

swal

ksan

dpe

destrian

sign

als

tohe

lpw

alke

rscr

oss

busy

stre

etsin

my

neig

hbor

hood

.11

.Th

ere

are

tree

sal

ong

the

stre

etsin

my

neig

hbor

hood

.12

.Th

ere

are

man

yin

ter-

estin

gth

ings

tolo

okat

whi

lew

alki

ngin

my

neig

hbor

hood

.13

.Th

ere

are

man

yat

-trac

tive

natu

rals

ight

sin

my

neig

hbor

hood

(suc

has

land

scap

ing,

view

s).

14Th

ere

are

attrac

tive

build

ings

/hom

esin

my

neig

hbor

hood

.Th

ere

isso

muc

htraffi

cal

ong

near

bystre

etsth

atit

mak

esit

diffi

cult

orun

plea

sant

tow

alk

inm

yne

igh-

borh

ood.

15.Th

esp

eed

oftraffi

con

the

stre

etIl

ive

onis

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

29

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

usua

llyslow

(30

mph

/50

kph

orle

ss).

16.M

ostdr

iver

sex

ceed

the

posted

spee

dlim

itsw

hile

driv

ing

inm

yne

ighb

orho

od.

Ther

eis

ahi

ghcr

ime

rate

inm

yne

ighb

or-

hood

.17

.Th

ecr

ime

rate

inm

yne

ighb

orho

odm

akes

itun

safe

togo

onw

alks

during

the

day.

18.Th

ecr

ime

rate

inm

yne

ighb

orho

odm

akes

itun

safe

togo

onw

alks

atni

ght.

19.Abo

utho

wlo

ngw

ould

itta

keto

wal

kfrom

your

hom

eto

the

near

estSu

perm

arke

t/Oth

erfo

od/g

roce

ry,

smal

lgro

cery

/con

ve-

nien

ce,f

ruit/

veg

mar

ket,

bake

ry,

butc

hersh

op/P

osto

f-fic

e/Any

scho

ol,e

le-

men

tary

,oth

er,n

ur-

sery

/Tra

nsit

stop

/Any

restau

rant

,fas

tfo

od,

non–

fast

food

,caf

é/co

ffee

plac

e/Pa

rk/

othe

rpu

blic

open

spac

e/Gym

/fitn

essfa

-ci

lity,

recr

eatio

nce

nter

,sw

imm

ing

pool

/Lib

rary

/Vid

eostor

e/Dru

gstor

e/ph

arm

acy/

Book

stor

e/Oth

ersh

opsan

dse

r-vi

ces

33Va

nCa

uwen

-be

rget

al.

(201

2)

Belg

ium

48,8

79Fl

emish

adul

tsag

edab

ove

65w

ithin

135

mun

icip

aliti

esco

llect

edin

2004

–201

0

surv

ey,t

heSt

udy

Serv

ice

ofth

eFl

emish

Gov

ernm

ent

1.Sh

ortd

ista

nces

tose

r-vi

ces;

2.Num

berof

shop

s3.

Acc

essto

Publ

ictran

spor

t

Nei

ghbo

rhoo

dFr

eque

ncy

ofcy

-cl

ing

fortran

spor

-ta

tion

SE/O

TNo

1.Num

berof

shop

s,sa

tisfa

ctio

nw

ithpu

blic

tran

spor

twas

positiv

ely

rela

ted

toda

ilycy

clin

gfo

rtran

spor

tatio

n.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

30

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

4.Pr

esen

ceof

publ

icto

ilets

5.Pr

esen

ceof

benc

hes

6.Pr

esen

ceof

cros

sing

s7.

Cond

ition

ofside

-w

alks

8.Abs

ence

ofhi

ghra

mps

9.Tr

affic

safe

ty10

.Fe

elin

gsof

unsa

fety

11.St

reet

light

ing

12.Abs

ence

ofde

cay

13.Abs

ence

ofno

ise

14.Gre

ener

y15

.Pu

blic

tran

spor

tatio

nsu

bscr

iptio

ns(%

ofol

derad

ults)

2.Fe

elin

gsof

unsa

fety

wer

ere

late

dto

ade

-cr

ease

dlik

ely-

hood

ofda

ilycy

clin

gfo

rtran

spor

tatio

nin

fe-

mal

esbu

tno

tin

mal

es.

3.Tr

affic

safe

tyw

asne

-ga

tivel

yre

late

dto

daily

cycl

ing

for

tran

spor

tatio

n.

34Ba

dlan

det

al.

(201

3)

Aus

tral

ia90

9ad

ults

who

mm

ovin

gin

to74

new

hous

ing

de-

velo

pmen

tsin

Perth

in20

03–2

004

1.Su

rvey

2.Obj

ectiv

een

vir-

onm

enta

ldat

a,3.

GIS

Perc

ieve

d:1.

Lots

ofgr

eene

ryin

the

neig

hbor

hood

2.In

tere

stin

gfe

atur

esin

the

neig

hbor

hood

3.Attra

ctiv

ebu

ildin

gsan

dho

mes

inth

ene

ighb

orho

od4.

Plea

sant

natu

ralf

ea-

ture

sin

the

neig

hbor

-ho

od5.

Pres

ence

ofw

alki

ngan

dcy

clin

gpa

thsin

the

neig

hbor

hood

6.Nei

ghbo

rhoo

dis

near

busy

road

s7.

Traffi

csp

eeds

are

slow

onne

arby

stre

ets

8.M

any

traffi

cslow

ing

devi

cesin

the

neig

h-bo

rhoo

d9.

Nei

ghbo

rhoo

dstre

ets

dono

tha

vem

any

culs-d

e-sa

c10

.Nei

ghbo

rhoo

dha

sm

any

four

-way

inte

r-se

ctio

ns11

.M

any

alte

rnat

ive

rout

esin

the

neig

h-bo

rhoo

dObj

ectiv

e:

Nei

ghbo

rhoo

dFr

eque

ncy

ofcy

-cl

ing

forre

crea

tion

and

tran

spor

tpe

rw

eek

SEYe

s1.

Resp

onde

ntsw

hope

rcei

ved

they

lived

near

busy

stre

ets

wer

eha

lfas

likel

yto

star

tcy

clin

gfo

rre

crea

tion

than

thos

eliv

ing

onqu

i-et

erro

ads.

2.Re

spon

dent

sliv

ing

inhi

gher

wal

kabl

ene

ighb

orho

ods(r

e-cr

eatio

n)or

with

high

erstre

etco

nnec

-tiv

ity(a

sm

easu

red

byGIS

)w

ere

mor

elik

ely

tostar

tcy

clin

gfo

rre

-cr

eatio

nw

hen

com

-pa

red

with

thei

rre

-fe

renc

egr

oups

.3.

No

sign

ifica

ntre

la-

tions

hips

betw

een

ob-

ject

ive

mea

sure

sof

the

neig

hbor

hood

and

upta

keof

cycl

ing

for

tran

spor

t.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

31

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

12.St

reet

conn

ectiv

ityz-

scor

e(m

edia

nsp

lit)

13.Re

side

ntia

lden

sity

z-sc

ore

(med

ian

split

)14

.La

ndus

em

ix(r

ecre

a-tio

n)z-

scor

e(m

edia

nsp

lit)

15.La

ndus

em

ix(tra

ns-

port)z-

scor

e(m

edia

nsp

lit)

16.W

alka

bilit

yin

dex

(re-

crea

tion)

z-sc

ore

(med

ian

split

)17

.W

alka

bilit

yin

dex

(tra

nspo

rt)z-

scor

e(m

edia

nsp

lit)

18.Cy

clab

lero

adra

tio(m

edia

nsp

lit)

19.Cy

cle

path

leng

thw

ithin

1600

mse

rvic

ear

ea(m

edia

nsp

lit)

35Be

enac

kers

etal

.(20

12)

Aus

tral

ia12

89ad

ults

who

mm

ovin

gin

to74

new

hous

ing

deve

lopm

ents

inPe

rth

in20

03–2

004

1.Nei

ghbo

rhoo

dEn

viro

nmen

tW

alka

bilit

ySc

ale

2.GIS

ina

1600

mbu

ffer:

1.Co

nnec

tivity

,2.

Reside

ntia

lden

sity

,3.

Land

-use

mix

,4.

Num

berof

destin

a-tio

nsre

leva

ntfo

rtran

spor

tor

recr

ea-

tion

5.Acc

essto

mix

edse

r-vi

ces—

scal

e6.

Nei

ghbo

rhoo

dae

sthe

-tic

s—sc

ale

7.Tr

affic

haza

rds—

scal

e8.

Maj

orba

rrie

rspr

esen

t9.

Loca

lpar

king

isdi

ffi-

cult

10.Acc

essto

park

11.Acc

essto

cycl

ing

path

s12

.Pe

destrian

cros

sing

spr

esen

t13

.Num

berof

tran

spor

tde

stin

atio

ns14

.Num

berof

recr

eatio

nde

stin

atio

ns

Nei

ghbo

rhoo

dDur

atio

ntim

e(m

ins)

ofcy

clin

gfo

rtran

spor

tan

dcy

clin

gfo

rre

crea

-tio

n

SEYe

s1.

That

grea

terob

jec-

tive

reside

ntia

lden

-sity

,inc

reas

edac

-ce

ssto

apa

rk,a

ndm

ore

recr

eatio

n-re

-la

ted

destin

atio

nsw

ere

positiv

ely

as-

soci

ated

with

anin

-cr

ease

intran

spor

t-re

late

dcy

clin

gaf

ter

relo

catio

n.2.

Ade

crea

sein

obje

c-tiv

eco

nnec

tivity

,in-

crea

sed

acce

ssto

ser-

vice

s,an

dm

ore

pede

strian

cros

sing

sw

ere

mar

gina

llyas

so-

ciat

edw

ithth

eup

take

oftran

spor

t-rel

ated

cycl

ing.

3.An

incr

ease

inob

jec-

tive

conn

ectiv

ityw

asas

soci

ated

with

the

upta

keof

recr

eatio

nal

cycl

ing.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

32

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

36M

aki-O

pas,

deM

unte

r,M

aas,

den

Her

tog,

and

Kuns

t(2

014)

Net

herlan

ds69

7ch

ildre

nag

ed10

–18

year

sin

Am

ster

dam

Stat

istic

sNet

herlan

dsan

dth

eDep

artm

entfo

rRe

sear

chan

dSt

atistic

sat

the

Mun

icip

ality

ofAm

ster

dam

2.GIS

3.NEW

Squ

estio

n-na

ire

surv

ey

1.Th

ere

are

man

yat

-trac

tive

natu

rals

ight

sin

my

neig

hbor

hood

(e.g

.lan

dsca

pe,

view

s),

2.M

yne

ighb

orho

odis

gene

rally

free

oflit

ter,

3.It

issa

feto

cycl

ein

orne

arm

yne

ighb

or-

hood

,4.

Ther

ear

ecy

clin

gpa

thsin

orne

arm

yne

ighb

orho

od5.

Ther

ear

em

any

maj

orin

ters

ectio

nsin

my

neig

hbor

hood

.6.

The

perc

enta

geof

gree

nar

eaw

ithin

ara

dius

of50

0m

arou

ndth

ere

side

nce.

7.Th

eEu

clid

ean

dis-

tanc

ebe

twee

nsc

hool

and

hom

e

Nei

ghbo

rhoo

d1.

Ave

rage

num

berof

days

,2.

Dur

atio

ntim

e(m

ins)

ofcy

-cl

ing

toan

dfrom

scho

olpe

rw

eek

SE/O

TNo

1.Fo

rev

ery

unit

ofin

crea

sein

the

bi-

cycl

e-frie

ndly

infra-

stru

ctur

e,th

ere

was

a21

%in

crea

sein

the

odds

forcy

clin

gto

and

from

scho

ol(b

icyc

le-fr

iend

lyin

-fras

truc

ture

in-

clud

es,c

ycle

path

sin

my

neig

hbor

-ho

od,t

raffi

cis

slow

,stre

etsar

ew

elll

itan

dth

ere

are

pe-

destrian

cros

sing

san

dtraffi

clig

hts,

the

traffi

cin

tens

ity.

37Zh

ao(2

013)

Chin

a61

3ho

useh

old

head

sof

613

hous

ehol

dsin

Beiji

ngfrom

60co

mm

uniti

es.

Surv

ey1.

Reside

ntspe

rhe

ctar

ein

built

-up

area

(att

hesu

b-di

strict

leve

l)(1

00pe

rson

s/ha

)2.

Jobs

perhe

ctar

ein

built

-up

area

(atth

esu

b-di

strict

leve

l)(1

00jo

bs/h

a)3.

Ajo

bs–h

ousing

bal-

ance

inde

xm

easu

red

atth

esu

b-di

strict

leve

l4.

Dista

nce

toth

eol

dci

tyce

nter

(Tia

nanm

ensq

uare

)from

the

cent

roid

ofco

mm

unity

(km

)5.

One

-way

com

mut

ing

time

(min

utes

)6.

Adi

vers

ityin

dex

mea

sure

dby

entrop

ym

etho

d7.

Leng

thof

loca

lstree

tspe

rsq

uare

kmin

Nei

ghbo

rhoo

dSh

are

ofcy

clin

gfo

rco

mm

utin

gSE

No

Hig

herde

stin

atio

nac

-ce

ssib

ility

,ahi

gher

num

berof

excl

usiv

ebi

-cy

cle

lane

s,a

mix

eden

-vi

ronm

entan

dgr

eate

rco

nnec

tivity

betw

een

loca

lstree

tste

ndto

in-

crea

seth

eus

eof

the

bicy

cle.

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

33

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

built

-up

area

with

inth

elo

cali

mpa

ctra

nge

(km

/squ

are

km)

8.Num

berof

cros

sing

sof

stre

etsw

ithin

the

loca

lim

pact

rang

e(u

nit)

9.Le

ngth

ofm

ain

road

and

expr

essw

ayspe

rkm

inbu

ilt-u

par

eaw

ithin

the

loca

lim

-pa

ctra

nge

(km

/sq

uare

km)

10.Num

berof

cros

sing

sof

mai

nro

adan

dex

-pr

essw

aysw

ithin

the

loca

lim

pact

rang

e(u

nit)

11.Le

ngth

ofbi

cycl

ela

nesse

para

ted

from

mot

oriz

edtraffi

cpe

rkm

inbu

ilt-u

par

eaw

ithin

the

loca

lim

-pa

ctra

nge

(km

)12

.Dista

nce

tocl

oses

tm

etro

stat

ion

from

the

cent

roid

ofco

mm

u-ni

ty(k

m)

38M

aet

al.

(201

4)

Am

eric

a83

0re

spon

sesin

thre

ene

ighb

orho

odsin

Portla

nd

the

Regi

onal

Land

Info

rmat

ion

System

(RLI

S)from

Portla

ndM

etro

,Re

fere

nce

USA

,

1.To

talm

ilesof

stripe

dbi

kela

nes,

mul

ti-us

epa

than

dlo

w-tr

affic

thro

ugh

stre

etsw

ithin

one

mile

ofho

me.

2.Num

berof

stre

etin

-te

rsec

tions

with

thre

eor

mor

eva

lenc

esdi

-vi

ded

byto

taln

umbe

rof

inte

rsec

tions

with

inon

em

ileof

hom

e3.

Num

berof

busine

sses

tabl

ishm

ents

with

inon

em

ileof

hom

e(b

ank,

restau

rant

,li-

brar

y,po

stoffi

ce,

groc

ery

stor

e,ph

ar-

mac

y,ba

rs,b

ooks

tore

,co

nven

ient

stor

e,

Nei

ghbo

rhoo

dFr

eque

ncy

ofcy

-cl

ing

SENo

1.A

neig

hbor

hood

with

conn

ecte

dstre

ets,

near

bybu

si-

ness

esta

blishm

ents,

and

low

-traffi

cstre

etsco

uld

mak

ere

side

ntsfe

elth

atbi

cycl

ing

inth

ene

ighb

orho

odis

easy

and

safe

,with

near

byde

stin

atio

ns.

2.Pe

rcep

tions

ofth

een

-vi

ronm

entha

vea

sig-

nific

antpo

sitiv

eas

so-

ciat

ion

with

bicy

clin

gbe

havi

or,i

ndic

atin

gth

atre

side

ntsw

hope

rcei

veth

eirne

igh-

borh

ood

asbi

keab

le

(con

tinue

don

next

page

)

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

34

Tabl

eA.1

(con

tinue

d)

Aut

hor

Coun

try

Sam

ple

Built

envi

ronm

ent

fact

orsda

taso

urce

Built

envi

ronm

entfa

ctor

sAna

lyze

dge

o-gr

aphi

cun

itsCy

clin

gm

etrics

Cont

rols

1Se

lf-se

lec-

tion

con-

trol

led

Resu

lts

fitne

ssce

nter

,the

ater

,an

dch

urch

)4.

Ratio

ofar

eaw

itha

slop

ele

ssth

an25

%w

ithin

one

mile

ofho

me.

5.Fo

rm

eto

ride

abi

-cy

cle

forda

ilytrav

elfrom

hom

ew

ould

beea

sy6.

Ikno

ww

here

safe

bike

rout

esar

ein

my

neig

hbor

hood

7.M

any

ofth

epl

aces

Ine

edto

gett

ore

gu-

larly

are

with

inbi

cy-

clin

gdi

stan

ceof

my

hom

e

actu

ally

bicy

cle

mor

eof

ten.

39Ve

rhoe

ven

etal

.(20

17)

Belg

ium

882

adol

esce

ntsag

ed12

–16

year

sfro

mFl

ande

rsSu

rvey

from

pa-

nora

mic

colo

rph

otog

raph

s

1.Se

para

tion

ofcy

cle

path

2.Ev

enne

ssof

cycl

epa

th3.

Spee

dlim

it4.

Spee

dbu

mp

5.Tr

affic

dens

ity6.

Am

ount

ofve

geta

tion

7.M

aint

enan

ce

Indi

vidu

alre

spon

-de

nt1.

Cycl

ing

rout

esch

oice

2.Num

bero

fday

spe

rw

eek

and

aver

age

daily

dura

tion

time

ofcy

clin

gto

variou

sde

sti-

natio

nsw

ithin

the

last

7da

ys

SEYe

s1.

Ado

lesc

ents'p

refe

r-en

ceto

cycl

efo

rtran

spor

tw

aspr

e-do

min

antly

dete

r-m

ined

byse

para

tion

ofcy

cle

path

,fol

-lo

wed

bysh

orte

rcy

clin

gdi

stan

cean

dco

-par

ticip

atio

nin

cycl

ing.

2.Hig

herpr

efer

ence

sw

ere

obse

rved

fora

sepa

ratio

nbe

twee

nth

ecy

cle

path

and

mot

oriz

edtraffi

cby

mea

nsof

ahe

dge

vers

usa

curb

,ver

susa

mar

ked

line

1 .AT

=Attitu

dina

lvar

iabl

es.

SE=

Soci

oeco

nom

icva

riab

les.

WE

=W

eath

erva

riab

les.

LE=

Leve

lofs

ervi

ceva

riab

les.

ST=

Stat

ion

variab

les.

OT

=Oth

erva

riab

les.

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

35

Tabl

eA.2

Char

acte

riza

tion

ofstud

iesby

the

envi

ronm

enta

lfac

tors

mea

sure

dm

etho

d,th

een

viro

nmen

tals

cale

,the

type

(s)of

cycl

ing.

Stud

yNo.

Envi

ronm

entat

trib

utes

mea

sure

dEn

viro

nmen

tala

ttribu

tessc

ale

Cycl

ing

type

Obj

ectiv

ePe

rcei

ved

Mic

roM

eso

Mac

roTr

ansp

ort

Com

mut

ing

Recr

eatio

nGen

eral

1X

XX

2X

XX

3X

XX

X4

XX

XX

X5

XX

XX

6X

XX

X7

XX

X8

XX

X9

XX

X10

XX

X11

XX

XX

12X

XX

X13

XX

XX

14X

XX

X15

XX

X16

XX

X17

XX

XX

18X

XX

X19

XX

X20

XX

X21

XX

X22

XX

XX

23X

XX

24X

XX

X25

XX

X26

XX

X27

XX

X28

XX

XX

29X

XX

30X

XX

31X

XX

32X

XX

33X

XX

X34

XX

XX

X35

XX

XX

X36

XX

XX

37X

XX

38X

XX

X39

XX

XTo

tal

2622

1620

423

98

10

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

36

App

endixB

Tabl

eB.

1El

astic

ityof

cycl

ing

fortran

spor

tatio

nw

ithre

spec

tto

dens

ity.

Stud

yNo.

Mod

elN

YX

e

9M

ultil

evel

logi

stic

regr

ession

3280

Bike

mod

ech

oice

Popu

latio

npe

rhe

ctar

e0.

0430

964

Bina

rylo

gistic

mod

el90

2Bi

kem

ode

choi

cefo

rut

ilita

rian

trip

s#

Reta

iljo

bsw

ithin

1/2-

mile

buffe

r−

0.01

242

4M

ultiv

aria

telin

earm

odel

902

Day

sof

utili

tarian

bicy

clin

g#

Reta

iljo

bsw

ithin

1/2-

mile

buffe

r0.

0310

9117

Biva

riat

ean

dm

ultiv

aria

telo

gistic

regr

ession

573

bike

mod

ech

oice

Built

cove

rage

(bui

ltco

ver/

built

lots)

−0.

0518

4

Tabl

eB.

2El

astic

ityof

cycl

ing

fortran

spor

tatio

nw

ithre

spec

tto

land

use

mix

Stud

yNo.

Mod

elN

YX

e

7lo

gistic

regr

ession

mod

el12

06bi

ketrip

spe

rpe

rson

land

use

mix

entrop

y0.

1392

167

logi

stic

regr

ession

mod

el12

06bi

ketrip

spe

rpe

rson

land

use

mix

entrop

y−

0.42

601

9M

ultil

evel

logi

stic

regr

ession

3280

bike

mod

ech

oice

land

use

mix

entrop

y0.

0268

36

Tabl

eB.

3El

astic

ityof

cycl

ing

fortran

spor

tatio

nw

ithre

spec

tto

acce

ssib

ility

ofno

n-re

side

ntia

ldes

tinat

ion

Stud

yNo.

Mod

elN

YX

e

33Lo

gistic

regr

ession

4887

9bi

ketrip

spe

rpe

rson

Num

berof

shop

s0.

2220

33

Tabl

eB.

4El

astic

ityof

cycl

ing

fortran

spor

tatio

nw

ithre

spec

tto

stre

et/r

oute

conn

ectiv

ity

Stud

yNo.

Mod

elN

YX

e

9M

ultil

evel

logi

stic

regr

ession

3280

bike

mod

ech

oice

rout

ein

ters

ectio

nde

nsity

(#in

ters

ectio

ns/h

a)0.

1479

449

Mul

tilev

ello

gistic

regr

ession

3280

bike

mod

ech

oice

orig

inin

ters

ectio

nde

nsity

(#in

ters

ectio

ns/h

a)0.

0550

499

Mul

tilev

ello

gistic

regr

ession

3280

bike

mod

ech

oice

destin

atio

nin

ters

ectio

nde

nsity

(#in

ters

ectio

ns/h

a)0.

1100

987

logi

stic

regr

ession

mod

el12

06bi

ketrip

spe

rpe

rson

Stre

etde

nsity

0.00

9799

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

37

Tabl

eB.

5El

astic

ityof

cycl

ing

fortran

spor

tatio

nw

ithre

spec

tto

open

spac

e/g

reen

spac

e

Stud

yNo.

Mod

elN

YX

e

10Lo

gistic

mod

el60

37bi

ketrip

spe

rpe

rson

Pres

ence

oftree

s0.

1163

58

Tabl

eB.

6El

astic

ityof

cycl

ing

fortran

spor

tatio

nw

ithre

spec

tto

aesthe

tics/

attrac

tiven

ess

Stud

yNo.

Mod

elN

YX

e

30Lo

gistic

regr

ession

4612

bike

trip

spe

rpe

rson

Plea

sant

envi

ronm

entto

cycl

e0

Tabl

eB.

7El

astic

ityof

cycl

ing

fortran

spor

tatio

nw

ithre

spec

tto

cycl

ing

faci

litie

s/cy

clin

gpa

ths

Stud

yNo.

Mod

elN

YX

e

9M

ultil

evel

logi

stic

regr

ession

3280

bike

mod

ech

oice

%ro

adne

twor

k:de

sign

ated

bike

rout

e0.

0590

1210

Logi

stic

mod

el60

37bi

ketrip

spe

rpe

rson

bicy

cle

lane

s−

0.00

877

4bi

nary

logi

stic

mod

el90

2bi

kem

ode

choi

cefo

rut

ilita

rian

trip

sM

ilesof

bike

lane

(GIS

)0.

1695

514

mul

tivar

iate

linea

rm

odel

902

Day

sof

utili

tarian

bicy

clin

gM

ilesof

off-stree

tbi

kepa

th(G

IS)

−0.

0487

730

Logi

stic

regr

ession

4612

bike

trip

spe

rpe

rson

Pres

ence

ofcy

cle

path

s0.

2026

8

Tabl

eB.

8El

astic

ityof

cycl

ing

fortran

spor

tatio

nw

ithre

spec

tto

cycl

ing

safe

ty

Stud

yNo.

Mod

elN

YX

e

7lo

gistic

regr

ession

mod

el12

06bi

ketrip

spe

rpe

rson

Traffi

csa

fety

num

ber

0.14

916

9M

ultil

evel

logi

stic

regr

ession

3280

bike

mod

ech

oice

Pres

ence

oftraffi

cca

lmin

gfe

atur

es0.

2236

3610

Logi

stic

mod

el60

37bi

ketrip

spe

rpe

rson

Traffi

cca

lmin

gfe

atur

es−

1.43

988

10Lo

gistic

mod

el60

37bi

ketrip

spe

rpe

rson

Spee

dlim

it≤

30km

/h0.

8549

3130

Logi

stic

regr

ession

4612

bike

trip

spe

rpe

rson

Low

traffi

csp

eed

0.27

9973

4bi

nary

logi

stic

mod

el90

2bi

kem

ode

choi

cefo

rut

ilita

rian

trip

sPe

rcei

ved

ther

ear

eoff

-stree

tbi

kepa

ths

0.16

6457

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

38

Tabl

eB.

9El

astic

ityof

cycl

ing

forco

mm

utin

gw

ithre

spec

tto

stre

et/r

oute

conn

ectiv

ity

Stud

yNo.

Mod

elN

YX

e

37m

ultin

omia

llog

itm

odel

613

bike

mod

ech

oice

Conn

ectio

nsof

loca

lstree

ts1.

0439

3828

Bina

ryan

dOrd

ered

Logi

stic

Regr

ession

2030

bike

mod

ech

oice

Origi

nin

ters

ectio

nde

nsity

0.12

1523

28Bi

nary

and

Ord

ered

Logi

stic

Regr

ession

2030

bike

mod

ech

oice

Des

tinat

ion

inte

rsec

tion

dens

ity0

Tabl

eB.

10El

astic

ityof

cycl

ing

forco

mm

utin

gw

ithre

spec

tto

cycl

ing

faci

litie

s/cy

clin

gpa

ths

Stud

yNo.

Mod

elN

YX

e

16M

ultip

lere

gres

sion

90of

the

100

larg

estU

.S.c

ities

Bike

com

mut

ers

bike

lane

spe

r10

0,00

0po

pula

tion

0.25

16M

ultip

lere

gres

sion

90of

the

100

larg

estU

.S.c

ities

Bike

com

mut

ers

bike

path

spe

r10

0,00

0po

pula

tion

0.09

137

mul

tinom

iall

ogit

mod

el61

3bi

kem

ode

choi

ceEx

clus

ive

bicy

cle

lane

s0.

5187

24

Tabl

eB.

11El

astic

ityof

cycl

ing

forge

nera

lwith

resp

ectto

open

spac

e/gr

een

spac

e

Stud

yNo.

Mod

elN

YX

e

8GLM

Mm

odel

2687

coun

tof

bicy

cle

Pct.

ofw

ater

body

in1-

mile

buffe

r0.

328

GLM

Mm

odel

2687

coun

tof

bicy

cle

Pct.

ofw

ater

body

in0.

5-m

ilebu

ffer

0.22

8GLM

Mm

odel

2687

coun

tof

bicy

cle

Pct.

ofw

ater

body

in0.

25-m

ilebu

ffer

0.27

Tabl

eB.

12El

astic

ityof

cycl

ing

forge

nera

lwith

resp

ectto

low

slop

e

Stud

yNo.

Mod

elN

YX

e

8GLM

Mm

odel

2687

coun

tof

bicy

cle

Pct.

ofstee

par

eain

1-m

ilebu

ffer

−0.

718

GLM

Mm

odel

2687

coun

tof

bicy

cle

Pct.

ofstee

par

eain

0.5-

mile

buffe

r−

0.02

8GLM

Mm

odel

2687

coun

tof

bicy

cle

Pct.

ofstee

par

eain

0.25

-mile

buffe

r−

0.05

Tabl

eB.

13El

astic

ityof

cycl

ing

forge

nera

lwith

resp

ectto

cycl

ing

faci

litie

s/cy

clin

gpa

ths

Stud

yNo.

Mod

elN

yx

e

8GLM

Mm

odel

2687

coun

tof

bicy

cle

Bike

lane

in1-

mile

buffe

r0.

888

GLM

Mm

odel

2687

coun

tof

bicy

cle

Bike

lane

in0.

5-m

ilebu

ffer

0.56

8GLM

Mm

odel

2687

coun

tof

bicy

cle

Bike

lane

in0.

25-m

ilebu

ffer

0.59

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

39

References

Aires, L., Pratt, M., Lobelo, F., Santos, R.M., Santos, M.P., Mota, J., 2011. Associations of cardiorespiratory fitness in children and adolescents with physical activity,active commuting to school, and screen time. J. Phys. Act. Health 8 (Suppl. 2), S198–S205.

Badland, H., Knuiman, M., Hooper, P., Giles-Corti, B., 2013. Socio-ecological predictors of the uptake of cycling for recreation and transport in adults: results from theRESIDE study. Prev. Med. 57 (4), 396–399.

Badrinarayanan, V., Kendall, A., Cipolla, R., 2017. SegNet: a deep convolutional encoder-decoder architecture for image segmentation. IEEE Trans. Pattern Anal.Mach. Intell. 39 (12), 2481–2495.

Bailey, D.P., Boddy, L.M., Savory, L.A., Denton, S.J., Kerr, C.J., 2012. Associations between cardiorespiratory fitness, physical activity and clustered cardiometabolicrisk in children and adolescents: the HAPPY study. Eur. J. Pediatr. 171 (9), 1317–1323.

Bartholomew, K., Ewing, R., 2009. Land use-transportation scenarios and future vehicle travel and land consumption: a meta-analysis. J. Am. Plan. Assoc. 75 (1),13–27.

Beaglehole, R., Bonita, R., Horton, R., Adams, C., Alleyne, G., Asaria, P., ... Alliance, N., 2011. Priority actions for the non-communicable disease crisis. Lancet 377(9775), 1438–1447.

Borenstein, M., Hedges, L.V., Higgins, J.P.T., Rothstein, H.R., 2009. Introduction to Meta-Analysis. Wiley, Chichester, UK.Brownson, R.C., Hoehner, C.M., Day, K., Forsyth, A., Sallis, J.F., 2009. Measuring the built environment for physical activity: state of the science. Am. J. Prev. Med. 36

(4 Suppl. l), 99–123.Buehler, R., Pucher, J., 2011. Cycling to work in 90 large American cities: new evidence on the role of bike paths and lanes. Transportation 39 (2), 409–432.Cai, Y., Zhu, X., Wu, X., 2017. Overweight, obesity, and screen-time viewing among Chinese school-aged children: national prevalence estimates from the 2016

Physical Activity and Fitness in China-The Youth Study. J. Sport Health Sci. 6 (4), 404–409.Cao, X., Mokhtarian, P.L., Handy, S.L., 2009. Examining the impacts of residential self‐selection on travel behaviour: a focus on empirical findings. Transp. Rev. 29 (3),

359–395.Cerin, E., Nathan, A., van Cauwenberg, J., Barnett, D.W., Barnett, A., Council on Environment and Physical Activity (CEPA), 2017. The neighbourhood physical

environment and active travel in older adults: a systematic review and meta-analysis. Int. J. Behav. Nutr. Phys. Act. 14.Cerin, E., Saelens, B.E., Sallis, J.F., Frank, L.D., 2006. Neighborhood environment walkability scale: validity and development of a short form. Med. Sci. Sport. Exerc.

38 (9), 1682–1691.Cervero, R., 2006. Alternative approaches to modeling the travel-demand impacts of smart growth. J. Am. Plan. Assoc. 72 (3), 285–295.Christiansen, L.B., Cerin, E., Badland, H., Kerr, J., Davey, R., Troelsen, J., ... Sallis, J.F., 2016. International comparisons of the associations between objective

measures of the built environment and transport-related walking and cycling: IPEN adult study. J. Transp. Health 3 (4), 467–478.Cole-Hunter, T., Donaire-Gonzalez, D., Curto, A., Ambros, A., Valentin, A., Garcia-Aymerich, J., ... Nieuwenhuijsen, M., 2015. Objective correlates and determinants of

bicycle commuting propensity in an urban environment. Transp. Res. D Transp. Environ. 40, 132–143.Davison, K.K., Werder, J.L., Lawson, C.T., 2008. Children's active commuting to school: current knowledge and future directions. Prev. Chronic Dis. 5 (3), A100.Day, K., 2016. Built environmental correlates of physical activity in China: a review. Prev. Med. Rep. 3, 303–316.DKS Associates, 2007. Assessment of Local Models and Tools for Analyzing Smart-Growth Strategies: Final Report. Retrieved from Sacramento, CA. https://trid.trb.

org/view/814672.Ewing, R., Cervero, R., 2010. Travel and the built environment. J. Am. Plan. Assoc. 76 (3), 265–294.Ferdinand, A.O., Sen, B., Rahurkar, S., Engler, S., Menachemi, N., 2012. The relationship between built environments and physical activity: a systematic review. Am. J.

Public Health 102 (10), E7–E13.Fraser, S.D., Lock, K., 2011. Cycling for transport and public health: a systematic review of the effect of the environment on cycling. Eur. J. Public Health 21 (6),

738–743.Ghekiere, A., Van Cauwenberg, J., Mertens, L., Clarys, P., de Geus, B., Cardon, G., ... Deforche, B., 2015. Assessing cycling-friendly environments for children: are

micro-environmental factors equally important across different street settings? Int. J. Behav. Nutr. Phys. Act. 12, 54.Gomez, L.F., Sarmiento, O.L., Parra, D.C., Schmid, T.L., Pratt, M., Jacoby, E., Pinzon, J.D., 2010. Characteristics of the built environment associated with leisure-time

physical activity among adults in Bogota, Colombia: a multilevel study. J. Phys. Act. Health 7 (Suppl. 2), S196–S203.Gotschi, T., Garrard, J., Giles-Corti, B., 2016. Cycling as a part of daily life: a review of health perspectives. Transp. Rev. 36 (1), 45–71.Haennel, R.G., Lemire, F., 2002. Physical activity to prevent cardiovascular disease - how much is enough? Can. Fam. Physician 48, 65–71.Handy, S.L., Boarnet, M.G., Ewing, R., Killingsworth, R.E., 2002. How the built environment affects physical activity - views from urban planning. Am. J. Prev. Med. 23

(2), 64–73.Heinen, E., van Wee, B., Maat, K., 2010. Commuting by bicycle: an overview of the literature. Transp. Rev. 30 (1), 59–96.Herlihy, D.V., 2011. Bicycle: the History. Yale University Press, United States.Hino, A.A.F., Reis, R.S., Sarmiento, O.L., Parra, D.C., Brownson, R.C., 2014. Built environment and physical activity for transportation in adults from Curitiba, Brazil. J.

Urban Health Bull. N. Y. Acad. Med. 91 (3), 446–462.Horner, S., Swarbrooke, J., 2005. Leisure Marketing: A Global Perspective. Elsevier Butterworth-Heinemann.James, P., Troped, P.J., Hart, J.E., Joshu, C.E., Colditz, G.A., Brownson, R.C., ... Laden, F., 2013. Urban Sprawl, physical activity, and body Mass index: nurses' health

study and nurses' health study 11. Am. J. Public Health 103 (2), 369–375.Johnston, R., 2004. The urban transportation planning process. In: Hanson, S., Guiliano, G. (Eds.), The Geography of Urban Transportation. Guilford, New York, NY,

pp. 115–140.Kerr, J., Emond, J.A., Badland, H., Reis, R., Sarmiento, O., Carlson, J., ... Natarajan, L., 2016. Perceived neighborhood environmental attributes associated with

walking and cycling for transport among adult residents of 17 cities in 12 countries: the IPEN study. Environ. Health Perspect. 124 (3), 290–298.Li, S.J., Liu, Y.Y., Zhang, J.J., 2011. Lose some, save some: obesity, automobile demand, and gasoline consumption. J. Environ. Econ. Manag. 61 (1), 52–66.Lu, Y., 2018. Using Google Street View to investigate the association between street greenery and physical activity. Landsc. Urban Plan., 103435.Lu, Y., Sarkar, C., Xiao, Y., 2018. The effect of street-level greenery on walking behavior: evidence from Hong Kong. Soc. Sci. Med. 208, 41–49.Lu, Y., Xiao, Y., Ye, Y., 2017. Urban density, diversity and design: is more always better for walking? A study from Hong Kong. Prev. Med. 103, S99–S103.Ma, L., Dill, J., 2015. Associations between the objective and perceived built environment and bicycling for transportation. J. Transp. Health 2 (2), 248–255.Maki-Opas, T.E., de Munter, J., Maas, J., den Hertog, F., Kunst, A.E., 2014. The association between physical environment and cycling to school among Turkish and

Moroccan adolescents in Amsterdam. Int. J. Public Health 59 (4), 629–636.Mavros, P., Austwick, M.Z., Smith, A.H., 2016. Geo-eeg: towards the use of EEG in the study of urban behaviour. Appl. Spat. Anal. Policy 9 (2), 191–212.Mueller, N., Rojas-Rueda, D., Cole-Hunter, T., de Nazelle, A., Dons, E., Gerike, R., ... Nieuwenhuijsen, M., 2015. Health impact assessment of active transportation: a

systematic review. Prev. Med. 76, 103–114.Muhs, C.D., Clifton, K.J., 2016. Do characteristics of walkable environments support bicycling? Toward a definition of bicycle-supported development. J. Transp. Land

Use 9 (2), 147–188.Munshi, T., 2016. Built environment and mode choice relationship for commute travel in the city of Rajkot, India. Transp. Res. D Transp. Environ. 44, 239–253.Naess, P., 2009. Residential self-selection and appropriate control variables in land use: travel studies. Transp. Rev. 29 (3), 293–324.Oja, P., Titze, S., Bauman, A., de Geus, B., Krenn, P., Reger-Nash, B., Kohlberger, T., 2011. Health benefits of cycling: a systematic review. Scand. J. Med. Sci. Sport. 21

(4), 496–509.Orstad, S.L., McDonough, M.H., Stapleton, S., Altincekic, C., Troped, P.J., 2017. A systematic review of agreement between perceived and objective neighborhood

environment measures and associations with physical activity outcomes. Environ. Behav. 49 (8), 904–932.Pucher, J., Buehler, R., Bassett, D.R., Dannenberg, A.L., 2010a. Walking and cycling to health: a Comparative analysis of city, state, and international data. Am. J.

Public Health 100 (10), 1986–1992.Pucher, J., Dill, J., Handy, S., 2010b. Infrastructure, programs, and policies to increase bicycling: an international review. Prev. Med. 50, S106–S125.Rissel, C., Watkins, G., 2014. Impact on cycling behavior and weight loss of a national cycling skills program (AustCycle) in Australia 2010-2013. J. Transp. Health 1

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

40

(2), 134–140.Rojas-Rueda, D., de Nazelle, A., Tainio, M., Nieuwenhuijsen, M.J., 2011. The health risks and benefits of cycling in urban environments compared with car use: health

impact assessment study. Br. Med. J. 343.Rojas-Rueda, D., de Nazelle, A., Teixido, O., Nieuwenhuijsen, M.J., 2013. Health impact assessment of increasing public transport and cycling use in Barcelona: a

morbidity and burden of disease approach. Prev. Med. 57 (5), 573–579.Rothstein, H.R., Sutton, A.J., Borenstein, M., 2005. Publication Bias in Meta‐analysis: Prevention, Assessment and Adjustments. John Wiley & Sons, Ltd, Chichester.Saelens, B.E., Handy, S.L., 2008. Built environment correlates of walking: a review. Med. Sci. Sport. Exerc. 40 (7 Suppl. l), S550–S566.Saelens, B.E., Sallis, J.F., Black, J., Chen, D., 2003a. Neighborhood-based differences in physical activity: an environment scale evaluation. Am. J. Public Health 93 (9),

1552–1558.Saelens, B.E., Sallis, J.F., Frank, L.D., 2003b. Environmental correlates of walking and cycling_ Findings from the transportation, urban design, and planning lit-

eratures. Soc. Behav. Med. 25 (2), 80–91.Sallis, J.F., 2002. Neighborhood Environment Walkability Scale. Retrieved from. http://sallis.ucsd.edu/measure_news.html.Selala, M.K., Musakwa, W., 2016. The potential of strava data to contribute in Non-Motorised Tansport (NMT) planning in Johannesburg. In: XXIII ISPRS Congress,

Commission II, vol. 41. pp. 587–594 (B2).Sharmin, S., Kamruzzaman, M., 2017. Association between the built environment and children's independent mobility: a meta-analytic review. J. Transp. Geogr. 61,

104–117.Snizek, B., Nielsen, T.A.S., Skov-Petersen, H., 2013. Mapping bicyclists' experiences in Copenhagen. J. Transp. Geogr. 30, 227–233.Su, M., Tan, Y.Y., Liu, Q.M., Ren, Y.J., Kawachi, I., Li, L.M., Lv, J., 2014. Association between perceived urban built environment attributes and leisure-time physical

activity among adults in Hangzhou, China. Prev. Med. 66, 60–64.Sun, B.D., Ermagun, A., Dan, B., 2017. Built environmental impacts on commuting mode choice and distance: evidence from Shanghai. Transp. Res. D Transp. Environ.

52, 441–453.Sun, Y.R., Du, Y.Y., Wang, Y., Zhuang, L.Y., 2017. Examining associations of environmental characteristics with recreational cycling behaviour by street-level strava

data. Int. J. Environ. Res. Public Health 14 (6), 644.Szeto, W.Y., Yang, L.C., Wong, R.C.P., Li, Y.C., Wong, S.C., 2017. Spatio-temporal travel characteristics of the elderly in an ageing society. Trav. Behav. Soc. 9, 10–20.Tilt, J.H., Unfried, T.M., Roca, B., 2007. Using objective and subjective measures of neighborhood greenness and accessible destinations for understanding walking

trips and BMI in Seattle, Washington. Am. J. Health Promot. 21 (4), 371–379.Ton, D., Duives, D.C., Cats, O., Hoogendoorn-Lanser, S., Hoogendoorn, S.P., 2019. Cycling or walking? Determinants of mode choice in The Netherlands. Transp. Res.

A Policy Pract. 123, 7–23.Train, K., 1986. Qualitative Choice Analysis: Theory, Econometrics, and an Application to Automobile Demand. The MIT Press, Cambridge, MA.Tsing Hua University Planning and Design Institution, & Mobike, 2017. Bike Sharing and the City - 2017 White Paper. Retrieved from Mobike: http://www.sohu.

com/a/133766880_585110.United Nations, 1999. Standard Country or Area Codes for Statistical Use (M49).Voss, C., Sandercock, G., 2010. Aerobic fitness and mode of travel to school in English schoolchildren. Med. Sci. Sport. Exerc. 42 (2), 281–287.Walters, J., Ewing, R., Allen, E., 2000. Adjusting computer modeling tools to capture effects of smart growth. Transp. Res. Rec. (1722), 17–26.Wang, R., Lu, Y., Zhang, J., Liu, P., Yao, Y., Liu, Y., 2019a. The relationship between visual enclosure for neighbourhood street walkability and elders’ mental health in

China: Using street view images. Journal of Transport & Health 13, 90–102.Wang, R., Yuan, Y., Liu, Y., Zhang, J., Liu, P., Lu, Y., Yao, Y., 2019b. Using street view data and machine learning to assess how perception of neighborhood safety

influences urban residents’ mental health. Health & Place 59, 102186.Wang, Y., Chau, C.K., Ng, W.Y., Leung, T.M., 2016. A review on the effects of physical built environment attributes on enhancing walking and cycling activity levels

within residential neighborhoods. Cities 50, 1–15.Zhang, M., 2004. The role of land use in travel mode choice - evidence from Boston and Hong Kong. J. Am. Plan. Assoc. 70 (3), 344–360.

Further Reading

Beenackers, M.A., Foster, S., Kamphuis, C.B., Titze, S., Divitini, M., Knuiman, M., ... Giles-Corti, B., 2012. Taking up cycling after residential relocation: builtenvironment factors. Am. J. Prev. Med. 42 (6), 610–615.

Cervero, R., Sarmiento, O.L., Jacoby, E., Gomez, L.F., Neiman, A., 2009. Influences of built environments on walking and cycling: lessons from bogotá. Int. J. Sustain.Transport. 3 (4), 203–226.

Chen, P., Zhou, J., Sun, F., 2017. Built environment determinants of bicycle volume: a longitudinal analysis. J. Transp. Land Use 10 (1).de Vries, S.I., Hopman-Rock, M., Bakker, I., Hirasing, R.A., van Mechelen, W., 2010. Built environmental correlates of walking and cycling in Dutch urban children:

results from the SPACE study. Int. J. Environ. Res. Public Health 7 (5), 2309–2324.Delfien Van Dyck, E.C., Conway, Terry L., De Bourdeaudhuij, Ilse, Owen, Neville, Kerr, Jacqueline, Cardon, Greet, Frank, Lawrence D., Saelens, Brian E., Sallis, James

F., 2012. Perceived neighborhood environmental attributes associated with adults' transport-related walking and cycling_ Findings from the USA, Australia andBelgium. Int. J. Behav. Nutr. Phys. Act. 9, 70–83.

Emma J Adams, A., Sahlqvist, Shannon, Bull, Fiona C., Ogilvie, David, on behalf of the iConnect consortium, 2013. Correlates of walking and cycling for transport andrecreation_ factor structure, reliability and behavioural associations of PENS. Int. J. Behav. Nutr. Phys. Act. 10, 87–101.

Forsyth, A., Oakes, J.M., 2014. Cycling, the built environment, and health: results of a Midwestern study. Int. J. Sustain. Transport. 9 (1), 49–58.Heesch, K.C., Giles-Corti, B., Turrell, G., 2014. Cycling for transport and recreation: associations with socio-economic position, environmental perceptions, and

psychological disposition. Prev. Med. 63, 29–35.Heesch, K.C., Giles-Corti, B., Turrell, G., 2015. Cycling for transport and recreation: associations with the socio-economic, natural and built environment. Health Place

36, 152–161.Ma, L., Dill, J., Mohr, C., 2014. The objective versus the perceived environment: what matters for bicycling? Transportation 41 (6), 1135–1152.Mertens, L., Compernolle, S., Deforche, B., Mackenbach, J.D., Lakerveld, J., Brug, J., ... Van Dyck, D., 2017. Built environmental correlates of cycling for transport

across Europe. Health Place 44, 35–42.Mertens, L., Compernolle, S., Gheysen, F., Deforche, B., Brug, J., Mackenbach, J.D., ... De Bourdeaudhuij, I., 2016. Perceived environmental correlates of cycling for

transport among adults in five regions of Europe. Obes. Rev. 17 (Suppl. 1), 53–61.Moran, M.R., Plaut, P., Baron Epel, O., 2015. Do children walk where they bike? Exploring built environment correlates of children's walking and bicycling. J. Transp.

Land Use.Nielsen, T.A.S., Olafsson, A.S., Carstensen, T.A., Skov-Petersen, H., 2013. Environmental correlates of cycling: evaluating urban form and location effects based on

Danish micro-data. Transp. Res. D Transp. Environ. 22, 40–44.Owen, N., De De Bourdeaudhuij, I., Sugiyama, T., Leslie, E., Cerin, E., Van Van Dyck, D., Bauman, A., 2010. Bicycle use for transport in an Australian and a Belgian

city: associations with built-environment attributes. J. Urban Health 87 (2), 189–198.Piatkowski, D.P., Marshall, W.E., 2015. Not all prospective bicyclists are created equal: the role of attitudes, socio-demographics, and the built environment in bicycle

commuting. Travel Behav. Soc. 2 (3), 166–173.Sallis, J.F., Conway, T.L., Dillon, L.I., Frank, L.D., Adams, M.A., Cain, K.L., Saelens, B.E., 2013. Environmental and demographic correlates of bicycling. Prev. Med. 57

(5), 456–460.Sylvia Titze, B.G.-C., Knuiman, Matthew W., Pikora, Terri J., Timperio, Anna, Bull, Fiona C., van Niel, Kimberly, 2010. Associations between intrapersonal and

neighborhood environmental characteristics and cycling for transport and recreation in adults_ baseline results from the RESIDE study. J. Phys. Act. Health 7,423–431.

Titze, S., Stronegger, W.J., Janschitz, S., Oja, P., 2008. Association of built-environment, social-environment and personal factors with bicycling as a mode oftransportation among Austrian city dwellers. Prev. Med. 47 (3), 252–259.

Van Cauwenberg, J., Clarys, P., De Bourdeaudhuij, I., Van Holle, V., Verte, D., De Witte, N., ... Deforche, B., 2012. Physical environmental factors related to walking

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

41

and cycling in older adults: the Belgian aging studies. BMC Public Health 12, 142.Van Dyck, D., Cardon, G., Deforche, B., Sallis, J.F., Owen, N., De Bourdeaudhuij, I., 2010. Neighborhood SES and walkability are related to physical activity behavior

in Belgian adults. Prev. Med. 50 (Suppl. 1), S74–S79.Van Holle, V., Van Cauwenberg, J., Deforche, B., Goubert, L., Maes, L., Nasar, J., ... De Bourdeaudhuij, I., 2014. Environmental invitingness for transport-related

cycling in middle-aged adults: a proof of concept study using photographs. Transp. Res. A Policy Pract. 69, 432–446.Verhoeven, H., Ghekiere, A., Van Cauwenberg, J., Van Dyck, D., De Bourdeaudhuij, I., Clarys, P., Deforche, B., 2017. Which physical and social environmental factors

are most important for adolescents' cycling for transport? An experimental study using manipulated photographs. Int. J. Behav. Nutr. Phys. Act. 14 (1), 108.Winters, M., Brauer, M., Setton, E.M., Teschke, K., 2010. Built environment influences on healthy transportation choices: bicycling versus driving. J. Urban Health 87

(6), 969–993.Winters, M., Davidson, G., Kao, D., Teschke, K., 2010. Motivators and deterrents of bicycling: comparing influences on decisions to ride. Transportation 38 (1),

153–168.Winters, M., Teschke, K., Brauer, M., Fuller, D., 2016. Bike Score(R): associations between urban bikeability and cycling behavior in 24 cities. Int. J. Behav. Nutr. Phys.

Act. 13, 18.Zahabi, S.A.H., Chang, A., Miranda-Moreno, L.F., Patterson, Z., 2016. Exploring the link between the neighborhood typologies, bicycle infrastructure and commuting

cycling over time and the potential impact on commuter GHG emissions. Transp. Res. D Transp. Environ. 47, 89–103.Zhao, P., 2013. The impact of the built environment on bicycle commuting: evidence from beijing. Urban Stud. 51 (5), 1019–1037.

Y. Yang, et al. Journal of Transport & Health 14 (2019) 100613

42