Electric BRT Readiness and Impacts in Athens, Greece: A Gradient Boosting-Based Decision Support Framework

Abstract

The integration of electric buses into urban transportation networks is a priority for policymakers aiming to promote sustainable public mobility. Among available technologies, electric Bus Rapid Transit (eBRT) systems offer an environmentally friendly and operationally effective alternative to conventional modes. This study introduces a Machine Learning Decision Support Framework designed to assess the feasibility of deploying eBRT systems in urban environments. Using a dataset of 28 routes in the Athens Metropolitan Area, the framework integrates diverse variables such as land use, population coverage, proximity to public transport, points of interest, road characteristics, and safety indicators. The XGBoost model demonstrated strong predictive performance, outperforming traditional approaches and highlighting the significance of points of interest, land use diversity, green spaces, and roadway infrastructure in forecasting travel times. Overall, the proposed framework provides urban planners and policymakers with a robust, data-driven tool for evaluating the practical and environmental viability of eBRT systems.