Analysis of urban vitality and its driving factors in Zhengzhou’s main urban area based on multi-source data and XGBoost

• Multi-scale, multi-dimensional data-fused quantification of urban vitality. • XGBoost-SHAP-driven community-level vitality audit of central Zhengzhou with planning recommendations. • The top three key factors are electricity consumption grid, population density, and PM2.5. • Electricity consumption grids and PM2.5 exhibit threshold and synergistic effects. Under the global urbanization background, urban vitality, a key indicator for measuring sustainable urban development and human settlement quality, is extremely important to detect from multi − dimensions and analyze its driving mechanisms. Most existing studies only focus on a single dimension or macro scale. In this study, we chose Zhengzhou’s main urban area, which is a national central city of China, as the research object. We integrated multi-source spatiotemporal data (including mobile signaling data, Baidu heatmaps, Point Of Interest (POI) data, NPP-VIIRS nighttime light, NDVI, etc.) at the community scale to build a comprehensive urban vitality evaluation model covering economic, social, cultural, and ecological dimensions. Using the XGBoost model combined with SHAP interpretability technology, we deeply analyzed the spatial distribution characteristics of urban vitality and the nonlinear mechanisms and interaction effects of its driving factors. Our findings are as follows: (1) Zhengzhou’s main urban area shows a “three − main − four − sub” spatial structure of vitality: economic vitality is highly concentrated in the central business district; social vitality is around university towns and transport hubs; cultural vitality is consistent with social vitality; and ecological vitality is dominant in peripheral areas. (2) Electricity consumption grid data (EC) is the core driving factor of urban vitality, and population density (PD), environmental quality (PM2.5), and recreational facility distribution (DOLF) also have significant impacts. (3) There are complex nonlinear relationships and threshold effects among the driving factors: when the EC value exceeds 0.84, its positive contribution to vitality significantly offsets the negative impact of increasing PM2.5 concentration; there are also remarkable synergistic enhancement effects between EC and PD, as well as between EC and DOLF. This study reveals the formation mechanisms of urban vitality and the complex interactions among its influencing factors at the micro scale, offering empirical evidence and planning insights at the community scale for cities worldwide to solve the “growth − sustainability − equity” trilemma.