Rentable Properties

A static, filterable list of 2BHK and 3BHK homes for rent that are within a 30-minute peak-traffic drive of a chosen point. Every listing needs four-wheeler parking and a move-in date by January 2027.

The listings are collected from vetted property portals and refreshed daily.

How it works

Layout and data are deliberately separated:

Path	Role
site/index.html	Page shell / structure
site/style.css	Design, responsive layout, light + dark themes
site/app.js	Fetches data/*.json and renders the map, cards, filters, search and sort in the browser
data/properties.json	The listings. This is the only file a routine refresh changes
data/sources.json	Portals that were evaluated: verdict, ratings (Google reviews, app stores, Trustpilot) and reasoning
data/isochrone.geojson	The 30-minute drive zones drawn on the map
data/localities.json	Locality → coordinate cache, used when a listing has no exact pin
scripts/collect_*.py	Per-portal collectors that write listings in the collection format
scripts/update_data.py	Validates, geolocates, adds drive times, de-duplicates and merges into data/properties.json
scripts/travel.py	The drive-time model
scripts/build.sh	Validates data and assembles site/ + data/ into _site/
.github/workflows/pages.yml	Publishes to GitHub Pages on every push to main
UPDATING.md	The refresh runbook followed by the scheduled job

Changing the data never requires touching the site. Changing the design never requires touching the data.

Drive-time model

The drive time comes from road distance, not straight-line distance, divided by the city's measured rush-hour speed. The speeds are from the TomTom Traffic Index 2025: 14.6 km/h in the evening peak and 17.7 km/h in the morning peak. The listed minutes use the slower evening figure. Road distance comes from OSRM, and the map zones are Valhalla isochrones of 7.3 km and 8.85 km by road, both from OpenStreetMap data.

Filters

Search text, south of the chosen point only (on by default), bedrooms (2/3 BHK, defaults to 3), max rent (defaults to ₹45,000), max drive time, property type, furnishing, move-in date, posted by (owner/agent), source portal, and an option to include listings that are no longer live. Sort by newest (the default), distance, rent or size.

Run locally

sh scripts/build.sh
python3 -m http.server -d _site 8000   # open http://localhost:8000

Refreshing data

python3 scripts/collect_nobroker.py /tmp/nobroker.json
python3 scripts/collect_olx.py /tmp/olx.json
python3 scripts/collect_magicbricks.py /tmp/magicbricks.json
python3 scripts/update_data.py merge /tmp/nobroker.json /tmp/olx.json /tmp/magicbricks.json --swept nobroker,olx,magicbricks

See UPDATING.md for the full procedure, including portals that are collected by hand.

Deployment

Every push to main that touches site/, data/ or the build scripts runs .github/workflows/pages.yml. The workflow builds _site/ and deploys it to GitHub Pages. You can also run it from the Actions tab.

Automated refresh

An AI agent runs the refresh on a daily schedule, following UPDATING.md. It pushes the updated data/properties.json to main, which triggers a redeploy. Always confirm a listing with the owner or agent before paying anything.