HomeRadar
HomeMapAdmin
Data dossier · 2026-07-12

Every dataset behind every answer.

HomeRadar is a synthesis layer over Singapore government open data — we don't scrape, we don't license, we don't make up numbers. Every figure on the map and in the AI answers is traceable to one of the sources below. This page lists what's loaded, where it came from, when it last refreshed, and what its known limitations are.

3.0K
private condos
URA
131K
sale transactions
URA · 5y rolling
369K
rental contracts
URA · 4q rolling
230K
HDB resale records
data.gov.sg · 2017-01+
113K
zoning polygons
URA Master Plan 2025
183
MRT + LRT stations
LTA DataMall
3.2K
amenities
OneMap themes
312
schools
MOE · OneMap
The roster — every dataset, sourced and dated

Eleven datasets, all government, all open, all updated automatically.

#01

URA private condo transactions

Urban Redevelopment Authority
URA Data Service · Free with API key
Refresh · Weekly · Mondays 3am
Source ↗
131Ksale transactions

Every private residential sale transaction in Singapore for the last 5 years. Includes contract date, price, area (sqm), floor range, and project name. We compute current PSF, 3-year and 5-year gains per project from these.

Known limits

URA's free API only goes back 5 years. Strata area only — no carpet/livable area. Unit numbers redacted (URA Realis paid feed has them). Transaction area is sqm, converted to sqft for PSF.

#02

URA rental contracts

Urban Redevelopment Authority
URA Data Service · Free with API key
Refresh · Weekly · Mondays 4am
Source ↗
369Klease contracts

Individual rental leases for private residential. Includes monthly rent, area, and bedroom count (87.5% coverage on bedrooms). We derive median rent, rental PSF, and gross yield per project, plus per-bedroom breakdowns.

Known limits

Bedroom count missing on ~12% of contracts — those are excluded from per-bedroom stats. Last 4 quarters fetched per pipeline run for efficiency.

#03

URA development pipeline supply

Urban Redevelopment Authority
URA Data Service · Free with API key
Refresh · Weekly · Mondays 6am
Source ↗
76projects in pipeline

Active development pipeline — private residential projects under construction or planned. We compute district-level supply pressure (pipeline units / existing private units) to identify oversupply risk.

Known limits

Supply pressure metric is restricted to districts with ≥2,000 private units (denominator threshold). Districts dominated by HDB/landed are NULL by design — the ratio would be misleading.

#04

URA Master Plan 2025 land-use zoning

Urban Redevelopment Authority via data.gov.sg
Open Data Licence · Free, attribution
Refresh · Annual · January 3 6am
Source ↗
113Kzoning polygons

Every land parcel in Singapore with its planned land-use category — residential, commercial, mixed-use, parks, reserve sites, business districts. We spatial-join each project to its zone and compute distance to nearest commercial parcel.

Known limits

WHITE designation (URA's flexible mixed-use category — Marina Bay, DUO Residences) is normalized to 'mixed' so AI's 'mixed-use' filter finds it. Source is in SVY21 projection, converted at load time.

#05

HDB resale prices

data.gov.sg
Open Data Licence · Free
Refresh · Monthly
Source ↗
230Kresale transactions

Every HDB flat resale transaction from 2017-01 to 2026-05. We compute rolling 12-month median PSF per planning area, then a per-condo "premium over local HDB" multiple — the canonical sanity check on whether a private condo is overpriced for its neighborhood.

Known limits

Premium % covers only 1175 of ~2,345 apartments (50.1%) — high-end CCR areas (Marina, Orchard, Tanglin, Newton) have no HDB blocks, so premium can't be computed. Landed properties excluded from premium calc — their PSF includes land area, making the ratio meaningless.

#06

Median household income by planning area

Singapore Department of Statistics · Census 2020
Open Data Licence · Free
Refresh · 10 years · next 2030
Source ↗
31planning areas

Median monthly household income by planning area, from the 2020 Census. Lets the AI answer 'condos in affluent areas' vs 'value neighborhoods' as more than vague vibe — a precise income-quartile filter.

Known limits

Census frequency is 10 years. 2020 data only; doesn't reflect post-pandemic shifts. Coverage: 31 of 55 planning areas — small/uninhabited zones excluded by SingStat.

#07

Planning area boundaries

data.gov.sg
Open Data Licence · Free
Refresh · Stable
Source ↗
55polygons

URA's 55 planning area polygons — the spatial unit we use for HDB premium, area income, and area-level overlays on the map. Every project's geographic context (Bukit Timah, Bishan, Punggol etc.) is resolved by spatial-joining its lat/lng to one of these.

#08

MRT and LRT stations

Land Transport Authority · DataMall
Singapore Open Data Licence · Free with API key
Refresh · Monthly
Source ↗
183stations (139 operational, 39 future)

Every operational MRT and LRT station, plus future stations from the Cross Island Line, Jurong Region Line, and TEL extensions. Station-to-condo distances are pre-computed via PostGIS and stored as nearest-MRT plus nearest-future-MRT per project.

Known limits

Future station coordinates are geocoded from LTA press release names via OneMap — accurate within 200–500m but not authoritative.

#09

Schools

OneMap (SLA) · Ministry of Education
OneMap Terms of Use · Free with token
Refresh · Monthly
Source ↗
312schools

Primary, secondary, JCs, and international schools. We resolve nearest schools by type per project to support 'family-oriented' searches.

#10

Amenities — hawkers, parks, gyms, libraries, clinics

OneMap (SLA) themes
OneMap Terms of Use · Free with token
Refresh · Monthly
Source ↗
3.2Kamenities

Hawker centres, parks, gyms, libraries, clinics, supermarkets, community clubs. Each project carries its nearest amenity per category, so 'walkable to a hawker' or 'near a park' isn't a guess.

#11

Building completion + management info

BCA via data.gov.sg (MCST register)
Open Data Licence · Free
Refresh · Monthly
Source ↗
2.0Kprojects matched

Building TOP date (≈ MCST constitution date), unit count, and management company per project. Used for 'building age' and unit-count overlays.

Known limits

2024 of 2994 projects matched (~68%). Older projects often pre-MCST registration.

Coverage at a glance

Not every project has every field. Here's where the gaps are.

We could fill gaps with imputed numbers and call it complete. We don't — missing data shows as "—" everywhere on the map and in the AI's answers, with reasons.

  • Planning areaEvery condo resolves to one of 55 planning areas via PostGIS spatial join100.0%
  • Zoning categoryEach project sits in a residential / mixed / commercial / park / reserve polygon99.5%
  • MRT lineMapped via LTA station code prefix (NS→NSL, DT→DTL, etc.)100.0%
  • Area median income31 of 55 planning areas — SingStat excludes uninhabited zones93.2%
  • District supply pressure25 of 28 districts — only districts with ≥2,000 private units98.3%
  • HDB premium % (apartments only)Limited by HDB block availability — high-end CCR areas (Marina, Orchard, Tanglin) have no HDB50.1%
What we don't have

Honest limitations.

Listing prices

We don't show what units are currently for sale. Closed transactions only — recent sale prices, not aspirational asking prices.

Unit-level detail

URA's free API redacts unit numbers. We can't show '#12-05 sold for $X' — only floor-range bands. URA Realis ($800/yr) has unit-level data; we don't have access.

Carpet area

URA reports strata area only — includes balcony, AC ledge, planter box. Singapore has no government source for carpet/livable area.

School rankings

MOE doesn't publish official PSLE/O-Level rankings as open data. We have school locations and types, but no quality scores. Third-party scrapes are unreliable.

Sales bedroom counts

URA sales API doesn't include bedroom count. We infer from per-project rental-data area→bedroom mappings (~85% accurate). Rental contracts have official bedroom counts (87.5% coverage).

Future MRT precision

Future station coordinates aren't published. We geocode station names from LTA press releases — accurate within 200–500m, but not authoritative.

HDB block coordinates

HDB resale data has block + street + town, no lat/lng. We aggregate to planning-area level for premium calc; individual blocks aren't on the map (yet — Phase 2 work to geocode all 9,704 blocks).

Price predictions

We don't forecast. Historical price trends and trajectories, yes. 'Will this go up?' — no, and any AI answer that pretends to is hallucinating.

How it stays fresh

Eight pipelines, all on cron.

Every dataset is refreshed by an Inngest job with retries, validation gates, and hash-based delta detection — if the source hasn't changed, we don't re-process. After every refresh, materialized views recompute distances, premiums, supply pressure, and all derived metrics.

  • Sale transactionsURA/TRANSACTIONSWeekly · Mon 3am
  • Rental contracts + mediansURA/RENTALSWeekly · Mon 4am
  • HDB resale pricesDATAGOV/HDB-RESALEWeekly · Mon 5am
  • Pipeline supplyURA/PIPELINE-SUPPLYWeekly · Mon 6am
  • Amenities + schoolsONEMAP/AMENITIES-SCHOOLSMonthly · 1st 5am
  • MRT stations + building infoLTA-BCA/MRT-MCSTMonthly · 1st 5am
  • Income by planning areaSINGSTAT/HOUSEHOLD-INCOMEYearly · Jan 2 6am
  • Master Plan zoningURA/MASTER-PLAN-ZONINGYearly · Jan 3 6am
HomeRadar
Sourced, not scraped. Updated, not invented.
Every figure on this page is reproducible from the cited sources. If a number looks wrong, it probably is — tell us.
← Back to home