Planning Future EV Charging Infrastructure by Forecasting Spatio-Temporal Adoption Trends Across Heterogeneous User Segments

The rapid transition to electric vehicles (EVs) requires a charging infrastructure that is both efficient and equitable. Conventional planning approaches, which often deploy chargers in proportion to current EV density, fail to account for the diverse characteristics of EV owners and the evolving patterns of adoption across different regions and time periods. This paper introduces an integrated, data-driven framework that addresses these limitations through three stages: segmentation of the EV market, spatio-temporal adoption forecasting for each segment, and optimizing charger placement through a constrained optimization model. The proposed optimization model incorporates equity constraints to ensure minimum service coverage for all user segments while maximizing overall utilization within a fixed budget. Methodologically, the paper contributes a transparent, reproducible framework that unifies user segmentation, geographically resolved adoption forecasting, and an equity-constrained MILP for charger placement. Applying this approach to a dataset of EV registrations in Washington State from 2010 to 2025 and extending it to projections through 2030 demonstrate important improvements in demand coverage. Overall coverage increases from 76.0% to 96.1% compared to a proportional-allocation baseline. More importantly, the proposed framework ensures a minimum of 70% coverage for all user segments. The presented approach is portable to other regions and budget scenarios. These findings show the potential for strategic, data-informed infrastructure planning that balances efficiency and equity, providing actionable insights for policymakers and network operators in the EV transition.