Predicting passenger crowding will allow train operators to manage the deployment of staff and rolling stock.
The latest available data from the Office of Rail and Road shows that passenger train use is up again, to 425.1million journeys in the second quarter of 2016/17. This figure is set to rise further.
The University of Kent’s research 'Development of intelligent predictive models for crowding on trains using data-driven methodologies' could help reduce crowding on trains, improving the customer experience and reducing costs associated with delays.
The model uses algorithms to predict where there will be crowding, taking into account regular passenger use and the effects caused by special events and disruption elsewhere on the network. It has already been tested using historical data from London Underground’s Victoria line and there are plans to test it on different parts of the network.
This project was one of the winners of the 'Data to improve the customer experience' competition, launched in 2015 by RSSB through the Rail Research UK Association (RRUKA), in association with the Rail Delivery Group. It shared the £220,000 funding with three other successful projects.
Associated article - RSSB funds research into new method of tailoring journeys