Revealing hidden patterns in transactional charging data: new insights from public charging stations in Brussels and San Francisco

The transition towards Electric Vehicles (EVs) has led many cities to develop networks of public charging stations. As a result, the observed charging data at these stations constitutes a valuable input to model charging demand. However, existing research that utilizes this data often relies on simple descriptive statistics (e.g. energy consumption or charging duration) that fail to capture the complex charging behaviour of EV drivers. In this study, we analyse two real-world charging datasets from Brussels and San Francisco and show how hidden behavioural patterns can be found. Our results indicate that groups of stations exist that are frequently visited by the same EV drivers. Furthermore, we are also able to reveal where these EV drivers divert when a preferred station is unavailable. Practitioners can use our methods to gain a better insight in how their infrastructure is used, and to more accurately determine where additional infrastructure is needed.