Abstract
We examined the nonlinear associations of the built environment with travel distances by driving, transit, and active travel and compared the relative contributions of local and regional built environment characteristics through applying gradient boosting decision trees to regional travel survey data in the Twin Cities, US. We identified the common thresholds of built environment characteristics associated with the three travel distances and inform planners of the efficient allocation of limited resources to planning efforts at different scales. We found prevalent threshold associations between built environment characteristics and travel distances. The thresholds suggest the common ranges of built environment characteristics that optimize the reduction of driving and the promotion of transit and active travel. For example, job accessibility should be larger than 800 thousand jobs within 20-min of driving and distance to downtown Minneapolis should be within 10 miles. The results also showed that regional characteristics collectively have a stronger influence on all three distances than local characteristics. The findings on transit and active travel differ from the common understanding in the literature. This study suggests that planners should pay more attention to metropolitan-scale planning and deploy programs that enhance regional accessibility.
Original language | English (US) |
---|---|
Article number | 103599 |
Journal | Journal of Transport Geography |
Volume | 109 |
DOIs | |
State | Published - May 2023 |
Bibliographical note
Publisher Copyright:© 2023 The Authors
Keywords
- Machine learning
- Nonlinear relationship
- Regional planning
- Threshold effect
- Travel behavior