Pattern recognition and spatial differentiation of built environment around urban sports facilities from perspective of facility hierarchy: a case study of Changsha city.

Under the context of healthy city development and the construction of 15-min life circles, the relationship between the built environment and residents' physical activity and health has attracted increasing attention. However, most existing studies adopt a residential-centered perspective, neglect the heterogeneity of service ranges across different levels of sports facilities, and insufficiently integrate objective built-environment characteristics with residents' facility-use behaviors and subjective perceptions. Taking Changsha city as a case study, this paper develops a hierarchical sports facility-based analytical framework for life circles and investigates the spatial differentiation and typological patterns of the surrounding built environment. Unlike fixed-buffer approaches, the framework employs level-specific measurement scales reflecting differences in potential facility service ranges and access modes. The study integrates multi-source geospatial data, including OpenStreetMap road networks, POI data, WorldPop population grids, and NDVI remote sensing data, to construct a multidimensional built environment indicator system covering accessibility, functional structure, facility supply, ecological quality, and population characteristics. Methods including GIS spatial analysis, one-way ANOVA, K-means clustering, and chi-square tests are employed to identify spatial patterns across different facility levels. Questionnaire results concerning activity-space choices, travel characteristics, facility use, and environmental perceptions are interpreted alongside the corresponding accessibility, functional, and ecological dimensions of the objective spatial analysis. This integration is conducted at the aggregate level and is not used to infer individual-level associations or causality. The results show that: (1) sports facilities in Changsha city exhibit a "large number of small-scale facilities and limited but spatially dominant large facilities" distribution pattern; (2) significant differences exist in the built environments surrounding different facility levels, with population density and POI functional diversity being the key distinguishing indicators; small facilities are mainly embedded in high-density residential areas, while large and extra-large facilities are more frequently located in areas with higher functional mix and higher mean NDVI values; (3) four typical built-environment patterns are identified, including high-density and service-intensive, low-intensity built-environment, high-NDVI and low-density, and functionally mixed and service-intensive types, showing a clear center-periphery gradient; (4) respondents mainly reported using parks, waterfront and mountain trails, and walking paths, and predominantly reached sports facilities through active travel within 15 min. These behavioral and perceptual results provide complementary context for the accessibility, functional, and ecological patterns identified by the spatial analysis, while concerns remain regarding pedestrian continuity and facility management.