Passenger counting and predictive models on the agenda for the 11th meeting
Passenger counting is a topic that has recently been thrown into the spotlight: in addition to ticketing data, it is an indispensable metric for traffic visibility, whether for network operation or customer relations.
The use cases based on passenger counting are varied and aim to improve the passenger experience on our networks:
- Passengers now expect up-to-the-minute ridership information. Keolis Orléans presented its project, that combines a predictive model with communicating ridership forecasts, to users via a mobile app. Keolis Lyon and the startup Affluence are also testing out a solution that aims to optimize the distribution of passengers along the metro platform, by indicating the doors to the least crowded cars in real time.
- Predictive models, while useful for many different businesses, have emerged as a great tool for adapting the available transport offer: Keolis Nederland has notably developed a predictive model based on AI that integrates factors influencing the use of transport (weather, events) in order to adapt its fleet. Keolis Côte Basque Adour is working with a local startup to include factors like climate and tourist seasons in its traffic forecasts.
- Keolis Besançon presented its "Fraudometer", which compares passenger counting and ticketing data in real time. It opted to use "nudges" to encourage passengers to validate their tickets. In this way, based on a non-validation threshold, the bus may announce a thank you message or one that encourages passengers to validate their tickets.