Map of Brisbane City Council paid parking meters with current-hour free-bay chance and a few high-signal historical charts.
brisbane-parking/
├── web/ # React + Vite + TypeScript frontend
├── api/
│ ├── py/ # Supported FastAPI runtime (ingest + server)
│ ├── src/ # Optional Rust/Axum reference implementation
│ ├── requirements.txt
│ └── parking.db # Generated SQLite (gitignored)
├── data/ # Canonical place for BCC CSVs (gitignored) + dataset notes
├── LICENSE # MIT (code)
└── NOTICE # BCC CC BY 4.0 data attribution
| Layer | Tech |
|---|---|
| Frontend | React, Vite, TypeScript, MapLibre, Recharts (web/) |
| API (supported) | Python FastAPI (api/py/) |
| API (optional) | Rust Axum (api/src/) — same routes/schema when your toolchain can build |
| Data | SQLite from BCC CSVs; 24h JSON cache under api/.cache/ |
Python is the supported way to run the API (especially on Windows where cargo link/build may be blocked). Keep Rust sources in sync as a reference, not as the default path.
Download CSVs from data.brisbane.qld.gov.au and place them under data/ (preferred):
| File | Dataset |
|---|---|
brisbane-parking-meters.csv |
brisbane-parking-meters |
parking-occupancy-forecasting.csv |
parking-occupancy-forecasting (current year) |
parking-occupancy-forecasting-2025.csv |
Same dataset, 2025 export if you use historical context |
Ingest also accepts the same filenames in the repo root as a fallback. Large CSVs and the SQLite DB are gitignored (~150MB+ of CSV). Dataset field notes live in data/docs/.
OCCUPANCY_PRED is a discrete 0–5 fullness score. Free chance = (5 - score) / 5.
Context features derived at ingest: weekday / weekend / QLD holiday, and season (DJF / MAM / JJA / SON).
Prerequisites: Python 3.11+, Node 20+.
# 1) CSVs in data/ (see above)
# 2) Ingest once (several minutes for ~6M occupancy rows)
make ingest
# or: cd api && python -m pip install -r requirements.txt && python -m py.ingest
# 3) Dev (API :8080 + Vite :5173 with /api proxy)
make dev
# or separately:
# make api
# make webRoot npm scripts also exist: npm run ingest, npm run dev:api, npm run dev:web.
cd api
# Needs a working MSVC or MinGW toolchain (+ Defender exclusion for target/ on Windows)
cargo run --bin ingest
cargo runGET /api/healthGET /api/meta— clock, suburb list, meter countGET /api/meters?west=&south=&east=&north=&suburb=&limit=250— pins for the visible map viewport (cached 24h per bbox)GET /api/meters/{meter_no}— chance, today/typical series, day-class context, nearby better optionsGET /api/carparks[?west=&south=&east=&north=]— off-street car parks (fromdata/carparks/carparks.json) with hours, height and grouped prices; no bbox returns all
The map only requests meters inside the visible bounds (debounced on pan/zoom). Zoom out past ~12.8 and pins unload until you zoom back in.
Current-hour lookup order: exact forecast → same zone/hour/day-class/season average → zone-hour average.
- Code: MIT — see
LICENSE - Data: © Brisbane City Council, CC BY 4.0 — see
NOTICE
Sources:
- https://data.brisbane.qld.gov.au/explore/dataset/brisbane-parking-meters
- https://data.brisbane.qld.gov.au/explore/dataset/parking-occupancy-forecasting
Production is Netlify for the frontend plus a remote FastAPI + SQLite API. Set VITE_API_BASE to the public API origin. See DEPLOY.md for the general steps.
Personal live URLs, host paths, and ops checklist: RodoVault Projects/Brisbane-Parking-Finder.md (outside this repo).
See CONTRIBUTING.md and CODE_OF_CONDUCT.md. Security reports: SECURITY.md.
Agents / AI assistants: start with AGENTS.md for a fast map of the repo, domain model, and which docs to open next.