Pinpointing available train seats

• The Swiss Federal Railways want to optimise the use of the existing infrastructure 

• A joint SBB and Zühlke team develops an algorithm based on machine learning 

• Thanks to the machine learning algorithm, SBB can better utilise its fleet and plan more precisely 

Finding available seats more quickly

Due to a growing population and greater mobility, Swiss Federal Railways (SBB) have capacity issues. They want to optimise use of their existing infrastructure. Until now, the only occupancy forecast was based on classes of travel. The SBB want to break this down by individual carriages. This would make it quicker to identify which seats are available and would mean that passengers are more evenly distributed across all carriages. Zühlke is contributing to the project with its expertise in project management and data science.  

Productive machine learning applications

A joint SBB and Zühlke team is developing an algorithm based on machine learning. Current data, such as feedback from the conductor, timetable information, group and individual seat reservations, public holiday dates, and weather forecasts are all being used. Zühlke is assisting with data processing, developing the algorithm and agile project management. 

Seat availability as a USP

The ability to predict which carriages have available seats is a unique selling point for the SBB. The machine learning algorithm allows the SBB to optimise use of its fleet and plan more accurately. Passengers can locate available seats much more quickly, and the overall user experience is enhanced. Train punctuality also improves. The solution is integrated into the online timetable, the SBB app, and the passenger information displays on platforms.