Android App for Local Train Status with Platform Alerts

Android App for Local Train Status with Platform Alerts is an Android app that answers one question for a daily local-train commuter: is my train on time, which platform will it arrive on, and has that changed since I last checked? Instead of a printed timetable, the app polls a live running-status feed, compares each train's actual position against its schedule, and shows delay in minutes alongside the expected platform, so a platform change is learned about before the train arrives rather than after it leaves. Saved commutes get delay push notifications checked in the background, and a GPS-based get-off alarm rings two stops before the destination station, even when the phone is locked. The demo runs on seeded simulated feed data; the production build connects to the documented live-feed integration point. Unlike a static offline timetable app, which can only estimate running status from the timetable, this project treats the live feed as the source of truth.

The flow is a loop, not a lookup: the commuter selects a train or a saved commute, and the app polls the live running-status feed for that route; each feed reply carries the train's last reported position and time, and the app compares it with the scheduled timetable to compute delay in minutes and read off the expected platform. A mismatch between the expected platform and the scheduled platform raises a platform-change alert, while the board refreshes on pull-to-refresh and on a timed poll. Saved commutes are checked in the background by WorkManager, which Android limits to a minimum 15-minute periodic interval, so delay alerts arrive without the app open. Separately, the GPS get-off alarm watches the phone's position against the destination station and rings two stops before it. The demo runs this whole loop on seeded simulated feed data; the production build points the same polling code at the documented live-feed integration point.

1. The user picks a home station and destination, or opens a saved commute; the app requests live running status from the feed for trains on that route.
2. The feed returns each train's last reported position and time; the app compares it with the scheduled timetable to compute delay in minutes.
3. The expected platform from the feed is shown next to the scheduled platform; a mismatch raises a platform-change alert in the app and as a notification.
4. The board refreshes on pull-to-refresh and on a timed poll; the demo re-renders a 6-train board in 0.1 ms median (demo-measured, 5 runs on the build machine).
5. Saved commutes are checked in the background by WorkManager at a minimum 15-minute periodic interval (Android-enforced limit), and delay, platform-change or cancellation alerts are pushed.
6. The GPS get-off alarm tracks distance to the destination station and rings two stops before it; the alarm state is stored locally and works offline once the trip is saved.

Tech stack:

• Kotlin · Jetpack Compose
• Android Studio
• WorkManager background scheduling
• Fused Location Provider (GPS)
• Retrofit · OkHttp (feed client)
• Room database (offline cache)
• Firebase Cloud Messaging (push alerts)
• Material 3 design

Live status is the centre of this app, and four capabilities carry it: the live running-status board shows delay in minutes and the expected platform for every train on the route; platform-change alerts fire the moment the feed reports a different platform from the scheduled one; a GPS-based get-off alarm rings two stops before the destination so a commuter never misses a stop while looking at a phone; and saved commutes are checked in the background so delay notifications arrive without opening the app. Around those four, train search and filter, next-train suggestions when a train is cancelled, and an offline cache of the last known status make the app usable on a crowded platform with patchy signal. Every feature below is implemented in the demo and the delivered source code.

• [Live running-status board] Each train shows scheduled time, expected time, delay in minutes, expected platform and a cancelled flag, computed from the live feed against the timetable — not from the timetable alone.
• [Platform-change alerts] When the feed reports an expected platform different from the scheduled one, the app raises an in-app banner and a push notification naming the old and new platform, as shown in the demo for train 90014 moving from Platform 2 to Platform 4.
• [GPS-based get-off alarm] The alarm tracks the phone's position against the destination station and rings two stops before it, even when the phone is locked; it works offline once the trip is saved.
• [Saved commutes with delay notifications] Daily routes are saved with a departure time; WorkManager checks the feed in the background and notifies the user of delays, platform changes or cancellations.
• [Train and station search] Filter the board by train number, train name or station so a commuter finds their train without scrolling the full departure list.
• [Next-train suggestions] When a saved train is cancelled, the app suggests the next departure on the same route with its platform, instead of leaving the commuter to re-search.
• [Offline trip cache] The last known status of saved trips is stored in a local Room database, so the app still shows the most recent delay and platform with its timestamp when the network drops.

• Complete Android source code (Kotlin + Jetpack Compose)
• Working demo build with seeded simulated feed data
• Feed-integration document (NTES integration point, polling design, error handling)
• Project report PDF, PPT presentation and viva Q&A document
• Setup instructions (Android Studio, SDK version, feed-key configuration)
• Project abstract PDF