A natural language search tool for managing Germany’s high-speed railway network
Together with Moritz Stefaner and Christian Laesser we have designed and built a new way to directly interact with vast operational databases at Deutsche Bahn long distance travel provider. It is named “Spotti” after the parent project it evolved from: Peak Spotting
Using simple natural language, Spotti allows train service operators to quickly get up-to-date information on train capacities, bookings and their predictions. Their millions of data points are instantly available to yield and traffic managers through a familiar search interface. This makes Spotti the primary tool for quick and direct tasks done by the teams working with Peak Spotting every day.
Spotti recognizes place names, train numbers, times and dates as well as several operational criteria such as estimated bookings. The pattern recognition allows to express queries to the database in flexible ways in order to support different search behaviors. The results themselves are fully working Peak Spotting interface elements, e.g. train lists, which provide familiar ways to further work with the selected trains.Investigating load balances of train operations with Spotti (obfuscated data).