Varying influences of the built environment on household travel in the United States – An update with 36 diverse regions and machine learning

People's daily travel behavior can have far-reaching impacts on issues ranging from climate change and public health to social equity. This study addresses a gap in the literature regarding the effects of built environments on travel behavior using a large (one million trips from 100,000 households), geographically precise (household XY coordinates), and multiregional (36 regions) dataset and machine learning analysis (boosted regression trees, BRT). BRT models were estimated in two stages for household VMT, walking, bicycling, and transit usage. The results show the built environment outweighs the impact of socioeconomics in all mode choice models and were significant in the trip generation models as well. Density and distance to transit were the most consistently influential built environment variables across the models estimated. Nonlinear and threshold effects were found for all dependent variables, suggesting key applications for planning. Suggested values are provided for minimizing VMT and maximizing active travel and transit usage.