既往研究主要聚焦于地方或区域模式，本研究则对美国本土48个州的公共电动汽车充电站（EVCS）可达性进行了全面的全国分析。通过整合空间统计方法（双变量局部Moran's I、共位分析）、机器学习（XGBoost）以及可解释的地理人工智能（GeoShapley），我们评估了EVCS分布与社会经济、人口统计、建成环境、政策及环境因素之间的关系。

While previous studies have largely focused on local or regional patterns, this study provides a comprehensive national analysis of public EVCS accessibility across the contiguous United States. Integrating spatial statistics (Bivariate Local Moran's I, colocation analysis) with machine learning (XGBoost) and explainable GeoAI (GeoShapley), we assess how EVCS distribution relates to socioeconomic, demographic, built environment, policy, and environmental factors.