{"package_name":"io.github.derweh.bayesianbahn","name":"BayesianBahn","summary":"Empirical arrival-time distributions for Deutsche Bahn trains","category":"Public Transport","icon_url":"/api/icon/io.github.derweh.bayesianbahn","latest_version_code":8,"latest_version_name":"0.4.0","apk_url":"/api/apk/io.github.derweh.bayesianbahn","apk_size":5264494,"apk_sha256":"60fd9dfc577670401e40b260647779a31e96fd975b075ec5a02ed957f5490134","source_kind":"fdroid-repo","repo_slug":"fdroid-main","last_updated":1789354847,"release_timestamp":1789210732,"description":"Early release: predictions are experimental — always cross-check\ntimes and connections with DB's official apps.\n\nBayesianBahn predicts when you will actually arrive — as a full\nprobability distribution, not a single number.\n\nEnter where you start, where you want to go and when (also future trips):\nthe app searches direct trains and journeys with one change — routes\nneeding two or more changes are not covered yet, so it will sometimes\nfind fewer connections than DB's own apps. For each option it shows:\n\n* the median predicted arrival time at your destination,\n* an 80% credible interval,\n* the full delay distribution as a chart,\n* the probability of catching each connecting train,\n* a Deutschland-Ticket filter that keeps you on regional trains.\n\nTransfers are propagated with Bayes' theorem: you board the first\nconnecting train that has not left yet — so a delayed earlier train\ncounts as catchable, and a missed connection honestly shifts the whole\ndistribution. Live station boards and per-train predictions (including\ncancellation rates) are also available.\n\nPredictions are empirical: they come from months of real historical runs\nof that exact train at that station (collected from Deutsche Bahn's\npublic IRIS API, CC BY 4.0), reweighted for recency and weekday. When\nDB reports an actual delay the forecast is anchored on it; when it\nreports nothing — which it does for most trains until shortly before\ndeparture — the history is left to speak for itself, which measurably\nbeats treating the timetable as a forecast. Delay history updates\nin-app — a small daily data release keeps predictions fresh to within a\nday. Trains without history get an honest Bayesian prior estimate.\n\nThe app talks only to the keyless public IRIS timetable endpoint and the\nproject's own data releases, needs no account, no API key, no Google\nservices, and collects no data.\n","categories":["Public Transport"]}