Citation

Chandrabose M, Rahim Mohammad Forkan A, Abe T, Owen N, Sugiyama T. Transp. Res. D Trans. Environ. 2023; 116: e103643.

Copyright

(Copyright © 2023, Elsevier Publishing)

DOI

10.1016/j.trd.2023.103643

PMID

unavailable

Abstract

Promoting physically active travel (walking + cycling) is a key strategy to enhance population health. It is important to understand for whom and where active and sedentary travel (car use) are common. However, little research has investigated how attributes of people and places interact in influencing these travel behaviors. Using travel survey data collected from 41,628 Australian adults, we employed decision tree modeling to untangle such complex relationships and to assess the relative importance of sociodemographic and environmental attributes in influencing active and sedentary travel durations. For active travel, attributes of places (suburb-level population density, distance to the nearest city center) were the most influential determinants. For car use, those two environmental attributes interacted with sociodemographic attributes (age, work status, household income) to form subgroups. Decision tree modeling appears to be effective to identify at-risk subgroups that may benefit from policy interventions.

Language: en

Keywords

Machine learning; Physical activity; Public health; Sedentary behavior; Travel survey