Station Utilisation Forecasting
Project information:
- Category: Software engineer internship
- Year: 2022
- Location: Evotrack, London, UK
My internship role is a software engineer focussing on data analysis and machine learning. My project encompassed a range of tasks, including leveraging Google Maps APIs for proximity analysis, generating diagnostic visuals for Paris charging station data, and deploying a station clustering algorithm.
K-means clustering of stations based on geometric location.
A pivotal focus was training a gradient boosting regressor, forecasting charging station usage using historical data and facility density. During this project, I have developed excellent programming skills and machine learning knowledge by exploring all challenging tasks independently.
On the left, utilisation factor of a station in a week is shown. Right, predicted and ground truth utilisation rate.