Data transparency · methodology

Where every
data point comes from

For every number and score card on the site: which source it comes from, how often it updates, its licence, and its known limits. All on one page.

Principle: all user-facing data is from official open data (gov.uk / ONS / police.uk / Land Registry / TfL / Ofsted / Environment Agency), clearly marked as wheretolive internal ML-derived, or listing and sold-price records collected from pages publicly displayed on Rightmove (listed separately below). We never blur “what we computed” vs “what the government published”.

Property listings + history

Live listing data (Rightmove)

103,493 listings · 1.64M photos

Active for-sale listings with asking_price / sq_ft / type / bedrooms / description / photos / tags

Update cadence
Collected automatically from adverts publicly displayed on Rightmove, twice daily (13:00 and 18:00 BST)
Licence / source
No licence from Rightmove; not affiliated with or endorsed by Rightmove. Listing text, photos and floorplans belong to the advertising agents, their photographers or Rightmove. Corrections and removals: [email protected]
Known limits
Portals don’t expose deeper data like EPC / planning · active listings only

HM Land Registry · Price Paid Data

~4.0M sales in London’s 33 boroughs since 1995

HM Land Registry’s record of residential sales in England and Wales that were sold for value and lodged for registration (price / date / type / new-build flag); we hold London

Update cadence
Monthly (HM Land Registry release, imported after)
Licence / source
Open Government Licence v3.0 · fully public
Known limits
Completion date only, no asking price / negotiation · the most recent months are still being registered · address_key occasionally mismatches on typos

Rightmove sold-price records (Value my home)

1.68M sale records for 1.08M homes

Sale records from sold-price pages publicly displayed on Rightmove: the same-postcode comparable sales in Value my home, and the data its price model was trained on

Update cadence
Collected automatically from sold-price pages publicly displayed on Rightmove, on a rolling basis since May 2026
Licence / source
No licence from Rightmove; not affiliated with or endorsed by Rightmove. Corrections and removals: [email protected]
Known limits
Used only in Value my home · covers the postcodes collected so far, not all of London

Internal delisted tracking

25,954 delisted / partially matched to LR

Listings found removed during our scrape · matched against LR to confirm a sale

Update cadence
Continuous · cloud delist queue
Licence / source
Derived from live listings + LR
Known limits
Match-score < 60 verdicts may be wrong (big address / type differences)

Internal price_history

83,324 price-change records

Snapshots of every asking_price change found between scrapes

Update cadence
Daily, following the scrape
Licence / source
Derived
Known limits
Price history before our first scrape is unavailable

Postcode / area scores

TfL · commute time + transit

UK-wide (live query API)

Public-transport + walking time to core CBDs / nearby stations / line types

Update cadence
Live at query time · cached 7 days
Licence / source
TfL Open Data Unified API
Known limits
Query rate limited by TfL API · occasionally falls back to default values

Ofsted · schools + ratings

~22K schools in England

Ofsted ratings (Outstanding/Good/RI/Inadequate) + type for schools within 1.5 km

Update cadence
Ofsted monthly update
Licence / source
gov.uk Open Data
Known limits
Hard 1.5 km radius cut-off · no private / international school admissions

police.uk · crime

Full England + Wales coverage

Crime types + monthly counts for the postcode’s police neighbourhood

Update cadence
Monthly (police.uk monthly release)
Licence / source
Open Government Licence
Known limits
Neighbourhood granularity only (≠ street level) · inconsistent reporting rates

Environment Agency · flooding

England

Long-term flood risk + historic flood events

Update cadence
Live at query time, cached 90 days
Licence / source
gov.uk Open Data
Known limits
Surface-water vs river-flooding models differ · postcode granularity limits

DEFRA · air quality + noise

Full UK coverage

NO₂ / PM2.5 / noise maps (road + rail + airport)

Update cadence
Annual (DEFRA annual update), cached 90 days
Licence / source
OGL
Known limits
1 km × 1 km cell granularity · doesn’t reflect hyper-local change (e.g. indoor air)

OSM (OpenStreetMap) / Natural England · parks

Full UK coverage

Greenspace / park count + area within 500m / 1km / 2km

Update cadence
Live at query time (OSM)
Licence / source
Open Database License (ODbL)
Known limits
OSM data quality varies by region · loose "park" vs "greenspace" definition

ONS · demographics

Full UK coverage (2021 Census)

Population structure / education / occupation mix / average income

Update cadence
Once a decade (Census) + ONS mid-year estimates
Licence / source
OGL
Known limits
2021 Census data · actual demographics may have shifted within 5 years

Property due-diligence data

EPC Register · energy certificates

~24M UK properties (public register)

EPC rating (A-G) / improvement tips / floor area / heating type

Update cadence
Live at query time (per-property)
Licence / source
OGL · gov.uk
Known limits
Some older homes have no EPC · we don’t query every listing

Valuation Office Agency · Council Tax

Full UK coverage

Council Tax band (A-H) + estimated annual

Update cadence
Live at query time
Licence / source
OGL
Known limits
Based on 1991 valuations (30 years out of date) · band-challenge logic in /council-tax-checker

Internal HMO licensing signal

Derived

HMO likelihood hint from bedrooms ≥ 3 + postcode demographics

Update cadence
At listing scrape time
Licence / source
Internal heuristic
Known limits
Not the council’s real HMO register · only a "may need a licence" hint

AI / ML derived data

SigLIP-2 luxury score

6,373,492 images scored (4,536,424 judged interior)

0-1 decor-tier score per image (zero-shot: 8 style anchors + 2 interior/exterior prompts)

Update cadence
Rolling append · ~50ms/image on an RTX 3060
Licence / source
Model: Google SigLIP-2 (Apache 2.0) · Scores: Derived
Known limits
Zero-shot, no fine-tune · edge-case accuracy not rigorously validated

Floorplan VLM parsing

On-demand · some historic backfill

Qwen3-VL-8B-AWQ floorplan → rooms / size / layout JSON

Update cadence
Phase 1 in progress
Licence / source
Model: Alibaba Qwen (Apache 2.0) · Output: Derived
Known limits
Bathrooms 71% — limited by fuzzy live-listing ground truth (inconsistent en-suite counting)

luxury_price_lookup · decor-price map

26,752 listings (14,324 with clean ppsf)

Listing-level luxury top3_mean + asking_price + price_per_sqft table

Update cadence
Rebuilt in batch after SigLIP scoring
Licence / source
Derived
Known limits
ppsf < £200 / > £3000 outliers filtered · uneven regional coverage (London-heavy)

External data (one-click links, not cached locally)

gov.uk Planning Permission

Council-specific portals

Planning permission records (build / convert / change of use)

Update cadence
Not cached · one-click jump to the relevant council portal
Licence / source
OGL
Known limits
No indexed planning DB · users go to the council site to search

Ofcom Broadband Checker

Full UK coverage

Actual fixed-line + mobile speeds

Update cadence
Not cached · /check links straight to the Ofcom checker
Licence / source
Ofcom Open Data
Known limits
Not in our local database

HM Land Registry title register

England and Wales

Title register + title plan (gov.uk £7 each to download)

Update cadence
Not cached · one-click link
Licence / source
Paid official copies from HM Land Registry (not open data)
Known limits
User pays £7 per document to buy direct from HM Land Registry

Data we don’t have

  • RICS Survey level structural data — a physical survey needs someone on site; commission a surveyor separately once an offer is accepted
  • Local Authority Search (CON29 / LLC1) statutory results — must go through a conveyancer via the council registration channel
  • Mortgage rate live feed — we use a 5.25% default (mid-2026 market level), not a live lender API
  • Planning permission indexed DB — only one-click links to council portals, nothing cached locally
  • On-the-ground street view / interior 360 / VR — use Google Street View or book a viewing
  • Privately held data (Sprift / Hometrack / PriceHubble and other B2B databases) — these are paid private datasets we don’t hold

Update cadence at a glance

Data classRefreshed
Live listingsTwice daily
Land RegistryMonthly
EPC / Council Tax / TfLLive at query time
Schools (Ofsted)Monthly
Crime (police.uk)Monthly
Air quality (DEFRA)Annual
SigLIP luxury scoresRolling append · live
Page updated · 2026-10-13 (Land Registry, Rightmove sold-price and coverage rows) · other counts as at 2026-06-30 · report issues or suggestions → /contact