In France, over 70% of daily commutes from suburbs to Paris are done using public transport, with SNCF Transilien operating most lines. But suburban rail networks are hard to optimize: trains overflow during peak hours and run mostly empty otherwise, leading to an inefficient use of the network, bottlenecks and money lost for operators, who want more balanced loads across trains.
In addition to that, commuters typically favor cars to public transport when they seek comfort, as they see public transport as a negative experience, limiting the opportunities for SNCF and increasing traffic on roads unnecessarily.
Lastly, access to data about passenger flow is expensive and hard to produce for operators: it is typically done via manual surveys, or by installing machines, neither of which being optimal.
Together with SNCF, Snips has built Tranquilien, a mobile app for iOS and Android giving prediction of available seats in trains up to a few hours in advance. It leverages a combination of machine learning to predict transport flow, and real-time user contributions to improve the model in case of unexpected events.
Furthermore, the app collected opt-in anonymous location traces, that were then aggregated into precise origin-destination matrices, at a fraction of the cost of deploying specialized hardware or doing manual surveys.
SNCF provided the historical data, transit APIs and marketing resources, while Snips provided the technology, app design and software engineering.
- 300K downloads, representing 40% of the addressable user base
- Over €1M in free PR resulting from the Snips-SNCF partnership
- 80% precision obtained by app data compared to surveys, while being 15x less expensive
- Comfort brought back to public transport for the first time!