london-live-2d

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

★ 11 Open GitHub ↗

Real-time 2D map of all London transport — live Tube, DLR, Overground, Elizabeth line, buses, trains, river boats, and more on one interactive map.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# London Live — 2D Real-Time Transport Map

**Every train, bus, boat, ship, plane and helicopter over London — live, on one map.**

### 🌍 Live demo: **[london.pengrubin.com](https://london.pengrubin.com)**

![Overview: central London with live tube, rail and river traffic](https://raw.githubusercontent.com/pengrubin/london-live-2d/main/docs/images/overview.png)

Inspired by [Zone One](https://london.jamespotter.dev/), rebuilt in 2D at
Greater-London scale. Author: **PENG**.

## What's on the map

| Layer | Source | How positions are derived |
|---|---|---|
| 🚇 Tube · DLR · Overground · Elizabeth (19 lines) | TfL Unified API | No GPS in the feed — positions **inferred from arrival countdowns**, scheduled inter-station run times, and animated along real OSM track geometry |
| 🚆 National Rail (431 stations) | Darwin via Rail Data Marketplace | Inferred from departure boards + calling points, pathed over a baked station-to-station rail graph |
| 🚌 Buses (~6,700 live) | DfT Bus Open Data (SIRI-VM) | Real GPS + per-vehicle α-β-style tracker; **self-learning route geometry** snaps buses to their true paths (auto-retrains daily from collected traces) |
| ⛴ Riverboats (RB1/RB4/RB6/Woolwich Ferry) | TfL | Countdown inference along the OSM Thames centreline, with curved pier approaches |
| 🚡 Cable Car · 🚊 Tram | TfL Unified API | Same countdown inference as the rail modes, animated along real OSM geometry |
| ⚓ Ships | AIS (aisstream.io) | Real positions; typed icons (cargo/tanker/passenger/tug); detail cards show dimensions, flag and a photo — ship photographs courtesy of [VesselFinder](https://www.vesselfinder.com), fetched per-MMSI |
| ✈️ Aircraft & helicopters | ADS-B (airplanes.live) | Real positions, dead-reckoned between polls; click for route lookup and an airframe photo via [Planespotters.net](https://www.planespotters.net) (with photographer attribution) |
| 📷 JamCams (878 traffic cameras) | TfL | Click for a live still, auto-refreshing |
| 🚧 Roadworks | TfL road disruptions | Live roadworks, closures and collisions as impact polygons + markers, refreshed every 2 min |
| 🌊 Tide gauges | Environment Agency | Live water levels along the tidal Thames (Teddington → Tilbury) |
| 🌧 Rain radar | RainViewer | Latest observed rainfall radar frame as a raster overlay |

Train detail cards also show a representative photo of the line's rolling stock,
bundled from [Wikimedia Commons](https://commons.wikimedia.org) (CC-licensed —
sources, licenses and authors in [docs/PHOTO_CREDITS.md](docs/PHOTO_CREDITS.md)).

![Click a train for its calling pattern, platform and delay status](https://raw.githubusercontent.com/pengrubin/london-live-2d/main/docs/images/train-detail.png)

## Using the map

**Control panel** (top-left, three tabs):

- **🏆 Board** — a live leaderboard ranking vehicles by cumulative distance travelled, over
  **Day / Week / Month**, with separate **Trains / Buses / Ships** boards. Double-click any row
  to fly to that vehicle (if it's currently running).
- **🚌 Filter** — type a bus route number (e.g. `24`) to spotlight it: matching buses stay
  **red** while every other bus **greys out but stays on the map and clickable**. Add several
  routes as chips. With the Buses overlay switched **off**, the filter instead shows *only* the
  matched routes and hides the rest.
- **🗺 Lines** — toggle any line or overlay on/off; **Select all / Unselect all** flips them
  all at once. Cable Car and Tram sit last (rarely used).

On phones the panel starts collapsed — tap the header to open it, and tapping the map collapses
it again so you can see the map.

**Hover** any vehicle, station or line for a quick tooltip (shared corridors list every line).
**Click** a vehicle for a detail card that follows it (destination, next-stop countdown, calling
pattern, platform, delay reason; name/speed for boats). **Click** a station for its live
departure board, zone, facilities, crowding, lift alerts and nearby cycle-hire docks.

## Icons at a glance

- **Trains** — a dot in the line's official colour (Central red, Victoria light-blue, …).
- **Buses** — a red route bullet oriented to its heading; **black** = parked; **grey** = greyed
  out by an active route filter.
- **Ships** — a hull icon coloured by AIS vessel type: 🔵 passenger · 🟡 cargo · 🔴 tanker ·
  🟢 fishing · 🟣 pleasure craft · ⚪ tug / pilot / SAR / unknown.
- **Aircraft** — a plane (helicopters too), pointing along its track.
- **📷 JamCam** cameras · **🚧 roadworks** · **🌊 tide gauges** — click any marker for details.

![Oriented bus bullets on street geometry](https://raw.githubusercontent.com/pengrubin/london-live-2d/main/docs/images/buses.png)
![AIS ships and traffic cameras along the Thames](https://raw.githubusercontent.com/pengrubin/london-live-2d/main/docs/images/ships-cams.png)

## Engineering highlights

- **Dirty-data hardening** for TfL's countdown feed: stale-"due" handoff, countdown-regression
  absorption (94 regressions observed in a 2-minute probe), monotonic track-space interpolation,
  branch hysteresis, coasting through feed flicker with destination-aware release.
- **All geometry is real**: tube tracks, rail corridors and the Thames centreline are stitched
  from OpenStreetMap; buses learn their routes from their own GPS traces (timetable shapes as prior,
  stop-anchored trimmed-mean averaging, quality-gated).
- **Polite API consumption**: batching + caching + per-upstream rate budgets keep TfL usage
  around 1% of the free allowance; all keys stay server-side.

## Running locally

```bash
# 1. basemap (~136 MB, not committed)
pmtiles extract https://build.protomaps.com/20260721.pmtiles data/london.pmtiles \
  --bbox=-0.55,51.25,0.35,51.72

# 2. keys — copy backend/.env.example to backend/.env and fill in:
#    TFL_APP_KEY (api-portal.tfl.gov.uk)      — required
#    AIS_API_KEY (aisstream.io)               — optional: ships
#    DARWIN_TOKEN (raildata.org.uk LDB key)   — optional: National Rail
#    BODS_API_KEY (data.bus-data.dft.gov.uk)  — optional: buses

# 3. run
cd backend && npm install && npm start        # :3000
cd frontend && npm install && npm run dev     # :5173

# 4. (one-off) bake static data if regenerating from scratch
node scripts/bake-routes.mjs && node scripts/bake-osm-geometry.mjs \
  && node scripts/bake-runtimes.mjs && node scripts/bake-nr-graph.mjs
```

## Layout

`frontend/` Vite + TypeScript + MapLibre GL · `backend/` Fastify proxies, pollers and the
self-scheduling bus-route learner · `scripts/` data baking (TfL, OSM, timetables, NR graph) ·
`data/` baked geometry (committed) + basemap/traces (not committed) · `docs/` architecture,
implementation plan, roadmap.

## Privacy

The "locate me" button uses your browser's geolocation only after you tap it (never on load).
Your coordinates are used purely to recentre the map in your browser — they are never sent to
the server, logged, or stored.

london-live-2d

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README