Technology
Technology and data
Follow a search from the native request to the local decision. Four detailed pages then document runtime, data access and model training down to individual metrics, and compare independent rail projects.
Path of a searchwithout a Bahnsparer rail-data server
- 01SearchRoute, travel day, travellers and BahnCard
- 02Request dataFares and journey data directly from the provider
- 03On your devicePrice comparison, reliability and price alert
- 04DecisionTravel day, connection, price alert or booking link
Four topics, four distinct questions
- Runtime and storageWhich code runs in the WebView and which runs natively? Where are routes, price alerts, keys and optional iCloud data stored?Open architecture
- Data sources and processingWhich sources provide fares, live data and carriage order? How does local request control limit calls and handle missing data?Open data sources
- Training and model qualityWhich historical months and features are used? Why were two tree models chosen, and how accurate are their probabilities?Open model documentation
- Other rail projects comparedWhich features are documented by BetterBahn, Besser Bahn, bahn.guru, FareWeave, Railwise and other projects? Where does Bahnsparer offer less?Open project comparison
Evidence for every claim
- Metrics
- The date range, sample and comparison method appear beside each metric.
- Data sources
- The primary sources and their roles appear directly beside the overview.
- Limitations
- Missing data, assumptions and untested cases remain explicitly named.