Impact of Charging Infrastructure Surroundings on Temporal Characteristics of Electric Vehicle Charging Sessions

In this paper, we apply a data-driven approach to analyse the temporal characteristics of charging sessions performed at a slow charging infrastructure. By using the variable selection ability of the Lasso method, combined with the bootstrap driven post-selection inference, we evaluate measures quantifying the potential impacts of charging infrastructure surroundings. We derive the description of the surroundings of the charging infrastructure from several publicly available datasets, representing social, demographic, business and physical environments. From the temporal characteristics, we focus on the average and standard deviation of the connection and charging time. We uncover a nonlinear relationship between the connection time and the charging time. The main driving factors behind the connection time are linked with the employment-related predictors and certain types of traffic influencing the variation of the connection time. The charging time is mainly affected by the economic wealth of residents. This study extends the knowledge about the electric vehicle driver charging behaviour and can be used to inform charging infrastructure deployment strategies.