Analysis and Prediction of Electric Vehicle Charging Supply and Demand Based on Data-Driven

The persistent problem of charging difficulties has long been a significant factor affecting users' decisions on choosing electric vehicles (EVs). Therefore, accurately grasping and applying the charging demand information is of paramount importance for every aspect of the electric vehicle ecosystem. This study based on a real dataset of EV charging, employs data analysis and mining techniques to dissect the charging demand information embedded within it. It defines the charging demands of EVs from three aspects, analyzes their spatiotemporal characteristics, and delves deeply into the charging supply and demand situations at different charging stations. Meanwhile, neural network models are utilized to predict users' charging demands. The research reveals various patterns of user charging behaviour at different charging stations, which can indicate whether the supply and demand at these stations are balanced. It also clarifies the rational configuration directions of charging facilities in different functional areas and enables effective predictions of future charging demands at various levels. This research provides multiple parties with insights into charging demands, contributing to the improvement of resource allocation efficiency, facilitating the construction of a more intelligent and efficient charging service system, and promoting the sustainable development of the electric vehicle industry.