Optimizing EV Charging Behavior: A Data-Driven Analysis of Local and Neighborhood Tariff Impacts

This paper proposes a data-driven methodology to analyze the impact of local and neighborhood tariffs on electric vehicle (EV) charging behavior. The study uses real-world data from public charging piles. It uses LSTM based dimensionality reduction, K-Means clustering based on cosine similarity and decision tree regression. It divides the charging stations into different clusters based on pricing model and occupancy. The results show that: (1) Occupancy rates show marked improvement when implementing dynamic pricing with competitively positioned local tariffs; (2) Rigid high-tariff frameworks persistently diminish utilization regardless of external pricing conditions; (3) The most deficient performance emerges in balanced tariff configurations (where local rates are marginally below adjacent stations yet both maintain static pricing). The proposed methodology is practical, offering actionable insights for optimizing pricing strategies and enhancing the efficiency of urban EV charging networks.