Train punctuality is usually high in Switzerland. When delays happen, railway staff and customers should be informed as soon as possible. Traditionally, delay forecasts are based on theoretical models and various assumptions - stemming back to an era when empirical data was hard to obtain. has started to challenge this approach: by looking for similarities with past situations, it predicts train delays using empirical data that is publicly available. All results are published in real time and can be compared to the official forecasts. The “open data” that is used has many restrictions: level of details and completeness are lower than what the companies possess internally. Nevertheless, in most cases, the empirical approach can predict delays earlier than the elaborated systems of the railway operators. Is data outperforming theory?


Time: 10:15 - 10:45

Track 5
organised by SSS

Speaker: Andreas Gutweniger