Prediction of Electric Vehicle Charging Stations Distribution Using Machine Learning

Electric vehicles (EV) are becoming more mainstream as consumer purchasing choices are in favor of this innovative technology. Hence, there is a need to install new electric vehicle charging stations (EVCS). Thus, determining where new stations should be installed is an apparent prospective challenge. In this paper, we tackle the mentioned challenge by creating a dataset enabling us to predict where new EVCS should be be installed. Machine learning models such as K-nearest neighbours, logistic regression, neural networks, and support vector machines can be applied to determine recommended EVCS installation places. City of Dubai located in United Arab Emirates (UAE) is selected to be the model of our case study. Population density, points of interest (POI), and the existence of security cameras are the features considered in this work. The performance of different machine learning models is evaluated using f1–score and accuracy metrics. Experiments show that the validation accuracy is 89% using K-Nearest Neighbors model. The dataset is publicly available at https://github.com/marwarai/Charging-Stations-Dubai.