Train delays are one of the reasons people can show up late to events or meetings. Until now, they seemed unpredictable. However, using the data from SBB we created Will I be late?, a website that can be used both to plan your journey and to predict possible delays in the trains based on the day that you are travelling, the route you are taking and the weather forecast of that day, among other factors.
How does it work?
Using the data provided by the SBB, we created a prediction system based on historical data of train delays. Does your journey involve passing through a city in which trains are frequently delayed? Are you travelling in a rainy day according to weather forecast? We take all of that into account to find the estimated delay of your train. The prediction system keeps getting better upon data collection.
How we built it
All the data processing was done using Python's Pandas. The website frontend was built with Angular 2, node.js for the server, whereas Python was used in the backend to both integrate the weather API to the website and to make the predictions from the data taken.
What's next for Will I be late?
Prediction systems are never perfect. Our system improves with the amount of data, but the algorithm can be improved in the future by knowing how exact are the predictions that it does.
Github commit hash: 095e164cdc06fee28d92051f91549d890dc5058a