Philippine Zonal Value Map

A free, public map of BIR residential zonal values (₱ per square meter of land) for Metro Manila and the provinces around it: Bulacan, Pampanga (with Angeles), Zambales (with Olongapo), Rizal, Cavite, Laguna, Batangas and Quezon (with Lucena). Bicol is next. Values are shown per barangay where the data allows. Every area carries a flag saying how good its number is.

Zonal values are BIR tax floors, not market prices. Read the map as a ranking. Check with the BIR Revenue District Office (RDO) before any transaction.

Data sources

What	Source	Where in repo
Zonal value schedules (primary)	BIR Department Orders, published per RDO at bir.gov.ph	raw/bir/RDO-<code>/*.xls(x)
Zonal values (fallback, until raw files exist)	BIR schedules as indexed by REN.PH	data/ncr_barangays.csv, data/pasay_barangays.txt, data/city_medians.csv
Barangay and city boundaries	PSA / NAMRIA, PSGC Dec 2023 release, via faeldon/philippines-json-maps	fetched into cache/ (not committed)

Quality flags

Every map feature has one of these flags:

• barangay_data: computed from that barangay's own schedule rows
• city_median: no barangay detail, so it shows the city's median residential value (hatched on the map)
• estimated: no data, so it shows the median of the neighbouring barangays it shares a border with (hatched and muted on the map)
• no_data: nothing usable (grey)

Values per barangay

• Interior lot: the lowest residential value in the barangay
• Prime: the highest residential value
• Typical: the median of street-level residential values. Where only a min and max exist, the geometric mean of the two is used instead.
• Adjusted: the value projected to the current year using the RDO's median growth between schedule revisions. Both the raw and adjusted values are always shown.

Repo layout

raw/bir/_inbox/     drop new BIR workbooks here; make data files them
raw/bir/RDO-<code>/ BIR workbooks filed by RDO, each holding every revision for that RDO
raw/bir/inventory.csv  one row per Department Order found in the workbooks
data/               fallback CSVs (REN.PH-indexed)
pipeline/           Python 3.11+ data pipeline
  aliases.csv       hand fixes: source barangay name -> PSGC name
  review/           generated: unmatched.csv, match report, parse errors
  tests/            pytest; fixtures/ holds sample BIR layouts
public/data/        pipeline output served by the site (zonal.geojson, cities.geojson, meta.json)
public/og.png       social preview, regenerated by `make data`
src/                Vite + TypeScript + MapLibre frontend
prototype/          v0 d3 prototype, kept for reference only

Rebuilding

Needs Node 20+ and Python 3.11+.

make venv    # one-time: creates .venv and installs pipeline/requirements.txt
make data    # fetches boundaries (cached), parses raw/ + data/, writes public/data/
make test    # pytest on the parser and name normalizer
npm install
npm run dev  # local site at http://localhost:5173

make data writes review files to pipeline/review/. If a barangay lands in unmatched.csv, add a row to pipeline/aliases.csv and run make data again.

Street and village values

Zoom in on any Metro Manila city, the Cavite and Laguna cities listed in streets.SOUTH, or any town in Batangas, Pampanga, Zambales or Quezon (streets.PROVINCE_SETS) to see BIR's street and subdivision values on the map: villages filled, streets as lines (dashed where BIR's stretch could not be pinned down, so the whole street is drawn). pipeline/streets.py matches each residential row of the current schedule to OpenStreetMap shapes in its own barangay; 62% of priced rows place. One cached OSM request per city (cache/osm/); one GeoJSON per city in public/data/streets/, loaded only when you zoom in. A city name used in two provinces (Rosario, San Juan) gets the province in its file name (rosario-batangas). Review files: pipeline/review/streets_<city>.csv. Method and limits: docs/pilot-muntinlupa-streets.md.

Other layers

Layer	Source	Pipeline
Roads & rail (current, under construction, proposed)	OpenStreetMap	pipeline/transport.py
Condominium values per building	BIR RC/CC rows on named OSM buildings (31% of rows place)	pipeline/condos.py, review pipeline/review/condos.csv
Flood hazard (share of barangay in medium/high hazard, 100-year rain)	UP NOAH / DREAM via LiPAD WFS (ODbL); NCR, Bulacan and Pampanga fully; Cavite, Laguna, Rizal and Batangas partly (Batangas 779 of 1,075 barangays; none for Mabini and Tingloy)	pipeline/factors.py flood
Elevation	Copernicus GLO-30 (surface model)	pipeline/factors.py elevation
Surface heat (dry-season afternoon land surface temperature)	Landsat 8/9 C2 L2 via Planetary Computer	pipeline/factors.py heat
Distance to rail station / expressway	OpenStreetMap	pipeline/factors.py access
Vacancy by submarket	hand-entered from broker reports: fill data/vacancy_submarkets.csv (with source and period); barangays per submarket in data/vacancy_submarket_barangays.csv. The option appears once a row has a number.	pipeline/build.py

python -m pipeline.factors all refreshes the factors (slow; downloads are cached in cache/factors/). After adding an area, python -m pipeline.factors <factor> --new computes a factor only for barangays that lack it, leaving existing values as they were.

Satellite imagery

The Satellite button puts imagery under the colours. With no configuration it uses Sentinel-2 cloudless 2016 by EOX (CC BY 4.0, 10 m per pixel: good for context, too coarse to see individual villages). For sharp imagery, add an Esri World Imagery key:

1. Create a free account at ArcGIS Location Platform (free tier: 2,000,000 basemap tiles a month; service stops rather than bills if exceeded, unless you enable pay-as-you-go).
2. Create an API key with the basemaps privilege only, and set its allowed referrers to https://ph-zonal-map.vercel.app (and http://localhost:5173 for development).
3. In Vercel: Project → Settings → Environment Variables, add VITE_ESRI_KEY with the key, then redeploy. Locally, put it in .env.local (see .env.example).

The key ends up in the browser by design; the referrer restriction is what keeps others from using it.

Adding new BIR schedules

1. Download the RDO's zonal value workbook from the BIR site and drop it, as is, into raw/bir/_inbox/ (subfolders and odd file names are fine).
2. Run make data. It reads each workbook's NOTICE sheet and Department Order headers, files it under raw/bir/RDO-<code>/ (for example RDO-027, RDO-53A), skips exact duplicates, and rewrites raw/bir/inventory.csv: one row per Department Order with its effectivity date and the cities it covers.
3. Check what needs a human:
   • raw/bir/inventory_log.txt: anything that needed judgment while reading the workbooks
   • pipeline/review/unmatched.csv: BIR barangay names that matched no PSGC barangay. Add a row to pipeline/aliases.csv and run make data again. Columns: city, source_name (as BIR writes it), psgc_name (a; b when one BIR block covers several barangays, (ignore) for sections that are not barangays), psgc_city (only when BIR filed it under another city), valid_before (only apply to schedules effective before this date, e.g. Bacoor's pre-merger names before 2023-07-29), note (evidence).
   • pipeline/review/plausibility.csv: values worth a second look
   • pipeline/review/parse_errors.csv.gz: every sheet row the parser could not use, with the reason (local only)
4. Commit and push; Vercel deploys main.

How values are computed

• Only residential land rows (class RR) count. Condominium rows (RC) are priced per sqm of floor area, not land, so they are left out.
• Each barangay uses the newest Department Order that covers it. Typical is the median of its street values, each street counted once; interior lot is the minimum, prime the maximum.
• If the newest schedule listing a barangay has no residential land row (BIR now classes it commercial or industrial only), older residential values are not used; it is treated as having no value.
• Where no BIR schedule matched a barangay, it is estimated as the median of the neighbouring barangays that have their own values (flagged estimated, hatched on the map). Only real values are used, never other estimates. With no such neighbour, it shows the city median (flagged city_median). A city with no values of its own (Pampanga today) is never estimated from a neighbouring city: it stays no_data.
• Adjusted to today. For each RDO, every street that appears in two consecutive revisions of the same barangay gives one annual growth rate; the RDO's rate is their median, taken from the transition into its current schedules. An RDO with fewer than 20 such pairs uses earlier transitions, then its province's median. Each value is then grown from the date it was actually set (a schedule that only repeats an older one's values does not count as a new date). Department Order averages are never compared. Rates per RDO are in pipeline/review/growth_by_rdo.csv; the summary is in docs/staleness-report.md.

License

Code: MIT. Data: BIR schedules are public records; boundary data is under the terms of its upstream repo.