Disparity in utilization rates of electric vehicle charging facilities : an empirical study of Hong Kong using machine learning

Abstract

This dissertation aims to provide insights for the planning and installation of public Electric Vehicle Charging Facilities (EVCFs) in Hong Kong, with a goal of facilitating the adoption of electric vehicles. The research objectives are to identify disparities in utilization rates of EVCFs, examine the locational factors affecting these rates using classical statistical multi-linear regression, and demonstrate the adoption of machine learning models in assisting future planning and installation of EVCFs through predictions. A quantitative approach is adopted, with classical statistical multi-linear regression used to identify the relationships between the utilization rate of EVCFs and locational factors, including demographic, socio-economic, and environmental factors of the vicinity. Machine learning non-linear regression models, such as tree-based models, are employed to generate predictions on the utilization rate of EVCFs. The findings provide insights for the planning and installation of public EVCFs, allowing resources to be allocated effectively to meet the demand of users and promote the adoption of electric vehicles in Hong Kong.