Decoding electric vehicle adoption using XGBoost and SHAP analysis

Abstract

As the shift towards sustainable transportation accelerates, it is vital to understand the factors driving electric vehicle (EV) adoption for effective policymaking. This study uses XGBoost and SHAP (SHapley Additive exPlanations) values to investigate the determinants of EV purchasing willingness in Florida, incorporating variables such as socioeconomic status, employer incentives, travel preferences, and spatial factors. The analysis shows that government subsidies have the most significant impact on EV adoption, with environmental concerns and travel preferences also playing key roles. The SHAP analysis identifies non-linear relationships with age, population density, and charger proximity. Additionally, approval of Tesla’s charging network openness and concern about climate change positively influence EV purchase willingness. This study enriches the literature by highlighting these complex, non-linear interactions and offers actionable insights for policymakers to enhance EV adoption. The key takeaways emphasize that accelerating EV adoption demands a comprehensive strategy involving enhanced financial incentives, increased environmental education, expanded charging infrastructure, and targeted approaches tailored to different age groups and geographic regions. These findings are crucial for developing targeted strategies to accelerate the transition to cleaner transportation solutions.