Citation

Luo Y, Liu Y, Tong Z, Wang N, Rao L. Travel Behav. Soc. 2023; ePub: ePub.

Copyright

(Copyright © 2023, Elsevier Publishing)

DOI

10.1016/j.tbs.2022.08.003

PMID

unavailable

Abstract

Rapid changes in the urban transportation environment amplify the impact of built environment factors on traffic collisions. Understanding the risk of potential harmful crashes for different populations as a result of changes in built environment thresholds is important for policy development and resource allocation. However, research on generic processes to effectively determine built environment threshold risks is lacking. Basing on extreme gradient boosting and SHapley Additive exPlanations, we establish a generic workflow for visualizing the risk thresholds for injurious traffic crashes for different genders and ages and built environment factors. The empirical results for the Wuhan metropolitan area (WMA) show that 1) thresholds for important population-specific variables require attention: bicycle speed for female children; street lighting for male children, female children, and female adults; and restaurant density for male children, female adults, and female seniors. 2) Threshold deviation shows the largest differences for the female children group. 3) Threshold risk maps show that high thresholds risk areas for male, adult, and child are located in the southeast of the WMA and those for females and elderly are located in the northwest. The generic workflow and empirical findings can help transportation planners develop specific practical strategies based on gender and age.

Language: en

Keywords

Built environment; Gender–age disparity; Injurious traffic crashes; Non-linearity; XGBoost model