Investigating the effects of urban morphology on vitality of community life circles using machine learning and geospatial approaches

In examining the associations between urban morphology and urban vitality, it is essential to focus on a scale that is directly observable, tangible, and appreciable in an individual's daily life. Past studies, however, have primarily concentrated on regional and urban scales. The complexities of human-environment interactions necessitate a more fine-grained measure of urban morphology beyond two-dimensional spaces. This study delves into the intricacies of horizontal and vertical urban surfaces within community life circles. It examines human-scale landscapes using a bottom-up metric system that includes eye-level quality, road network design, buildings and functions, and neighborhood location. The methods of deep learning and street view images are utilized to evaluate the eye-level morphology. The study then analyzes the heterogeneous impacts and interaction effects of morphological variables on urban vitality at the community life circle level through multiscale geographic weighted regression and geographical detector model. The findings reveal a core-periphery structure and strong spatial autocorrelations between the morphological variables and urban vitality. Furthermore, the determinants of urban vitality exhibit a blend of global, regional, and local spatial contexts. The human-observed greenery, sky openness, and public bus accessibility demonstrate significant spatial heterogeneity linked to local vitality, whereas street betweenness and commercial facilities exhibit limited effects. The interaction effects between morphological variables can either enhance or weaken their individual impacts, highlighting the need for community life circle planning that considers the combined effects of these variables. This study provides new insights into urban design approaches for promoting revitalization and micro-renewal at the human scale.