Differences in the influence of the built environment on physical activities for people with normal weight, overweight, and obesity.

The link between the built environment and physical activity (PA) has received substantial research attention in health geography and public health studies. Limited research efforts have been spent to investigate if the link would be different for people with or without obesity, and prior studies often overlook potential nonlinear associations. Knowledge of such potential differences will be essential for developing tailor-made urban planning or spatial policies to promote physical activities and health for different population groups. This study aims to address this research gap by exploring the nonlinear relationships between the built environment and PA engagement for people with normal weight, overweight, and obesity. Data are derived from a two-day household activity-travel diary survey conducted in 2018 in Shanghai, China. Using gradient boosting decision tree models, this study found distinct influences of the built environment on total PA duration and active PA for different weight status groups. Residents with obesity tend to have a shorter total PA duration and a lower rate of participating in active PA. By comparing the model results, it shows that the impact of the built environment on active PA is larger than that on the total PA duration for all three weight groups, and this increase is more pronounced for the obesity group. Population density, public transit, and land use mix are the three key built environment variables that exert nonlinear effects on PA engagement across all weight groups, with variations in trend and threshold values between groups. This research presents new insights into the relationship between the built environment and physical activities, which are relevant for promoting physical activities among people with different weight statuses.