Predicting Optimal Placement of Electric Vehicle Charge Stations Using Machine Learning: A Case Study in Glasgow, UK

Given the reduced costs of establishing Electric Vehicle (EV) charging stations and the need to improve their accessibility, it is essential for EV developers and governments to identify optimized locations for new stations. This study presents a predictive modeling approach using Machine Learning to determine the optimal sites for EV charging stations in Glasgow. While various machine learning methods are available for this purpose, we focus on Linear Regression due to its simplicity and high accuracy. By analyzing a comprehensive dataset of 207 AC charging stations-including factors latitude, longitude, population density, number of kinds of plugs, total number of plugs, price per kWh, traffic conditions, distance to the nearest station, Air Quality Index (AQI), and land value - we aim to provide urban planners and policymakers with actionable insights. The study evaluates the performance of the Linear Regression model in predicting the most suitable locations for charging stations. The results indicate that the proposed model achieved impressive accuracy, with R2, Mean Absolute Error (MAE), and Mean Squared Error (MSE) values of 0.89, 0.05, and 0.08, respectively. This research contributes to the ongoing efforts in urban planning for sustainable mobility by offering a data-driven methodology for EV infrastructure development.