Presyo

The problem

The Philippine government publishes a lot of price data: palengke prices from the Department of Agriculture, pump prices from the DOE, suggested retail prices from the DTI, exchange rates from the BSP. But it comes as PDFs and spreadsheets spread over a dozen websites, in different formats, with many reports a week.

Nobody checks five PDFs before going to the market. So the data exists, but most people never see it.

What I built

Presyo puts it all in one app: pick a city and see today’s prices for rice, meat, fish, vegetables, gasoline and diesel by brand, LPG, electricity rates, a Meralco bill estimate by kWh, water rates, jeepney and bus fares, and peso exchange rates.

Each price shows where it came from and when it was last updated, so you can check it against the original report.

How it works

• Ingest (Python, uv, httpx, pdfplumber): fetches each source and parses the PDFs and CSVs into one normalized format, keeping the history in the repository. DA and DOE tables differ by region and change format, so each source has its own parser with pytest tests. The BSP site rejects ordinary scripts, so it uses a browser-like TLS client.
• Places (shapely): reports write “Sta. Cruz”, “City of Cebu” or “Taguig”. A normalizer folds abbreviations and city prefixes and matches them against 1,736 PSGC places, with centroids computed from boundary polygons.
• Build: precomputes small JSON files per source and place, with headline items and sparklines, so the app never filters large files.
• App (Expo / React Native): resolves the nearest place on the device and fetches one small file per source. Hosting is static, so there is no server to run.
• Automation: GitHub Actions runs test → fetch → commit → build → deploy at 12:00 and 20:00 Manila time. Failures are logged as warnings and the build still ships with the data that did arrive.

Try it

The app is live. Use it in the phone frame on this page, or open presyo.pages.dev on your phone.