TY - JOUR
T1 - How the built environment shapes our daily journeys
T2 - A nonlinear exploration of home and work environments’ relationship with active travel in Shanghai, China
AU - Jiang, Huaxiong
AU - Zhang, Qinran
AU - Guo, Kaifei
AU - Helbich, Marco
AU - Yang, Haoran
N1 - Publisher Copyright:
© 2025 Elsevier Ltd
PY - 2025/2
Y1 - 2025/2
N2 - In this study, we examined the built environment's nonlinear relationships and threshold effects on the active travel of 1,002 full-time employees in Shanghai, China. Using gradient boosting decision trees (GBDTs), we considered active travel and objectively measured and perceived built environmental correlates at both the residential and workplace levels. Our results showed that 1) both residential and workplace built environment attributes tend to have a nonlinear relationship and that threshold effects exist until active travel is related; 2) compared with the perceived built environment, the objective features of the built environment had distinct explanatory power, while the relationships followed either (inverted) V-shapes or were roughly linear; 3) relationships of the built environment, regardless of whether they were measured objectively or perceived, deviated between residential and workplace locations; and 4) the built environment was more strongly related with walking rather than cycling at both the residential and workplace levels. Overall, the findings provide policy-makers with more effective neighborhood planning strategies to facilitate active travel-friendly environments.
AB - In this study, we examined the built environment's nonlinear relationships and threshold effects on the active travel of 1,002 full-time employees in Shanghai, China. Using gradient boosting decision trees (GBDTs), we considered active travel and objectively measured and perceived built environmental correlates at both the residential and workplace levels. Our results showed that 1) both residential and workplace built environment attributes tend to have a nonlinear relationship and that threshold effects exist until active travel is related; 2) compared with the perceived built environment, the objective features of the built environment had distinct explanatory power, while the relationships followed either (inverted) V-shapes or were roughly linear; 3) relationships of the built environment, regardless of whether they were measured objectively or perceived, deviated between residential and workplace locations; and 4) the built environment was more strongly related with walking rather than cycling at both the residential and workplace levels. Overall, the findings provide policy-makers with more effective neighborhood planning strategies to facilitate active travel-friendly environments.
KW - Built environment
KW - Cycling
KW - Machine learning
KW - Nonlinear relationships
KW - Walking
UR - https://www.scopus.com/pages/publications/85216292125
U2 - 10.1016/j.tra.2025.104377
DO - 10.1016/j.tra.2025.104377
M3 - 文章
AN - SCOPUS:85216292125
SN - 0965-8564
VL - 192
JO - Transportation Research Part A: Policy and Practice
JF - Transportation Research Part A: Policy and Practice
M1 - 104377
ER -