Spatiotemporal association of urban park characteristics and physical activity levels based on GTWR: A serial cross-sectional observational study

Physical inactivity poses a significant public health challenge. Numerous studies have argued that urban parks can promote physical activity (PA). However, evidence of physical activity observed over long periods of time in parks is still very lacking. Our serial cross-sectional study within urban parks in Shanghai investigated seasonal and daily variations in PA. Employing a geographically and temporally weighted regression (GTWR) model, we established spatiotemporal relationships between urban park characteristics and PA levels. We found that (1) Highest PA levels occur in winter, while the lowest levels are observed in summer, with a V-shaped daily pattern. (2) Robust and significant associations between specific park characteristics and PA levels, while other characteristics exhibited varying associations in different temporal and spatial contexts. These insights offer valuable guidance for urban planners and park designers aiming to create more active and sustainable urban environments.