Enhance Journey Planner with Predictive Travel Information for Smart City Routing Services

Route planning in public transport receives an increasing interest in smart cities and particularly in metropolitan cities where crowded and jammed traffic is daily recorded in the transportation network. The availability of digital footprints, such as ticketing logs, or load on board the trains, provides a relevant opportunity to develop innovative decision-making tools for urban routing of passengers in order to assist them to better plan their journeys. In this paper, we propose to enrich existing journey planners with predictive travel information to enhance the passenger travel experience during his journey. For that purpose, we augment the planned trips with predictive passenger flow indicators such as the load on board trains, and passenger attendees at the station. These indicators are forecasted along the journey with the help of the developed machine learning models. The experiments are conducted on a real historical dataset covering the Paris Region with a focus on a railway transit network that serves mainly the suburb of Paris.