Urban vitality reflects a city’s capacity to support economic activities, social interactions, cultural participation, and ecological sustainability. This study presents an urban vitality evaluation system encompassing cultural, economic, social, and ecological dimensions, while considering Taiyuan City (China) as the research area. Objective and subjective variables and multisource data were used to present an evaluation framework for the drivers of urban vitality. Getis–Ord Gi hotspot analysis and Local Moran’s I were used to characterize the spatial clustering patterns of multidimensional urban vitality. The categorical boosting-Shapley additive explanations (CatBoost-SHAP) method was employed to explain the impact of various influencing factors on urban vitality. The spatial heterogeneity and feature–location interactions of key influencing factors across different urban spaces were analyzed using the GeoShapley method; k-means clustering was employed to classify multidimensional vitality states and GeoShapley-based local driving mechanisms. The results revealed that social, economic, and cultural vitality exhibited H–H clustering in the central urban area, while ecological vitality was distributed in the northern and southern peripheral areas of the city. Building height (BH), Housing price (PR), land-use mix (LUM), Building views ratio (BV), POI diversity index (PDI), and building density (BD) were the primary influencing factors. BH, Sky view ratio (SV), BV, PR and LUM exhibited significant spatial heterogeneity. This study presents novel insights for policymakers to develop targeted policies for precise urban planning.