How varun1323/jaunt-route-recommender works
Jaunt is a Manhattan walking-route recommender: travelers choose landmarks and express preferences, then the app retrieves cached pedestrian routes, ranks them, and presents route details and geometry. The backend also exposes landmark data and can obtain a missing baseline through Valhalla. Geographic research and candidate access-point work support the route data, but the sampled excerpts do not establish a complete production pipeline between those utilities and the app. The graph reflects documented or excerpted connections; some UI rendering and data wiring is omitted where excerpts do not show it.
Main entry points: App.tsx App state · app.py Jaunt API · recommendation_engine.py Recommendation engine · fetch_routes.py Route fetcher
Components 23
Traveler experience
- App state
frontend/src/App.tsx - Journey selector
frontend/src/components/JourneySelector.tsx - Preference input
frontend/src/components/PreferenceInput.tsx - Recommendation card
frontend/src/components/RecommendationCard.tsx - Route list
frontend/src/components/RouteList.tsx - Route map
frontend/src/components/Map.tsx - Header
frontend/src/components/Header.tsx - Limitation notice
frontend/src/components/LimitationBanner.tsx - Error notice
frontend/src/components/ErrorBanner.tsx - API client
frontend/src/services/api.ts
Recommendation API
- Jaunt API
backend/app.py - Recommendation engine
backend/scripts/recommendation_engine.py - Preference parser
backend/scripts/preference_parser.py
Route data
- Landmark records
backend/data/landmarks.json - Route GeoJSON cache
backend/data/routes - Route fetcher
backend/scripts/fetch_routes.py
Geographic research
- Building collection
backend/scripts/collect_nyc_buildings.py - Geodata collection
backend/scripts/collect_geodata.py - Geometry utilities
backend/scripts/geodata_geometry.py - Access-point matching
backend/scripts/match_access_points.py - Access candidates
backend/data/pedestrian_access_point_candidates.json
Other components
- Traveler
- Valhalla routing
Connections 23
- Traveler → App state: uses
README.md - App state → API client: requests data
frontend/src/App.tsx - API client → Jaunt API: calls API
README.md - Jaunt API → Landmark records: reads
backend/app.py - Jaunt API → Route GeoJSON cache: loads routes
README.md - Jaunt API → Recommendation engine: requests ranking
backend/app.py - Recommendation engine → Preference parser: parses preference
backend/scripts/recommendation_engine.py - Recommendation engine → Route GeoJSON cache: scores routes
backend/scripts/recommendation_engine.py - Jaunt API → Route fetcher: generates missing route
README.md - Route fetcher → Valhalla routing: requests pedestrian route
README.md - Route fetcher → Route GeoJSON cache: writes GeoJSON
README.md - App state → Journey selector: renders
frontend/src/App.tsx - App state → Header: renders
frontend/src/App.tsx - App state → Limitation notice: renders
frontend/src/App.tsx - App state → Recommendation card: renders
frontend/src/App.tsx - App state → Route list: renders
frontend/src/App.tsx - App state → Route map: renders
frontend/src/App.tsx - App state → Error notice: renders
frontend/src/App.tsx - App state → Preference input: renders
frontend/src/App.tsx - Traveler → Building collection: researches places
- Traveler → Geodata collection: researches geography
- Geometry utilities → Access-point matching: supports matching
- Access-point matching → Access candidates: produces candidates