UK Listed Building Data

The WRENN Report assembles every public dataset that touches a UK listed building — the registers themselves, planning applications, Energy Performance Certificates, Land Registry sale prices, flood and radon risk, and architect attributions — and links them together at the building level. Everything below is free to search at /search.

Listed buildings — all four nations

493,181 listed buildings, scraped from the official heritage registers of all four UK nations and consolidated into a single search interface.

• England — 379,840 buildings (Historic England)
• Scotland — 67,463 buildings (Historic Environment Scotland)
• Wales — 30,116 buildings (Cadw)
• Northern Ireland — 15,762 buildings (NIEA)

By statutory grade: Grade I (13,029) · Grade II* (45,783) · Grade II (391,010). Buildings span 379 Local Planning Authorities.

429,856 buildings (87%) have had their official register entry rewritten into clean, modern English by a language model — abbreviations like "C18" expanded to "18th century", run-on prose restructured, factual content preserved. The original register text is also retained.

Planning applications and Listed Building Consent

883,305 listed-building-related planning applications, ingested daily from PlanIt, parsed for date received, decision status, applicant agent, and described works. 218,324 buildings (44%) have at least one matched application linked to them. Matching uses a multi-stage address pipeline: postcode + house number + street, with LLM verification for ambiguous cases.

Every application has been processed through GPT-4o-mini to extract structured "physical attribute" tags (extension, roof, glazing, solar, change of use, internal alteration, demolition, etc.) so you can search by what the work actually involves, not just by free text.

Energy Performance Certificates

32,876,616 domestic EPCs from the Department for Levelling Up, Housing & Communities (England & Wales) and statistics.gov.scot (Scotland), plus 1,632,981 non-domestic EPCs covering pubs, churches, museums, schools, and offices. Refreshed monthly (E&W) and quarterly (Scotland).

118,817 listed buildings have at least one EPC matched to them — the gap to total EPC volume reflects that most EPC certificates are for non-listed properties. Listed buildings skew strongly toward the lower energy bands: D and E are most common, only ~1.6% are rated B or above (versus ~15% nationally), and ~7% are rated G — consistent with solid uninsulated walls, single glazing, and oil heating in rural settings.

Northern Ireland is not currently covered for EPC. The data is not published as open data and would require an EIR request to the Department of Finance.

HM Land Registry sale prices

31,092,167 residential property transactions in England & Wales from January 1995 onwards, from HM Land Registry's Price Paid Data (Open Government Licence). 141,969 listed buildings have at least one sale matched to them — multiple sales per building are preserved so you can see a full price history.

Environmental risk

Every building (where the underlying flood/radon data is openly licensed) has a flood-zone classification and a radon-risk classification:

• Flood (477,419 buildings) — Environment Agency Flood Map for Planning (England), Natural Resources Wales (Wales), SEPA v3 Flood Maps (Scotland). Northern Ireland data is commercially licensed and not included.
• Radon (493,159 buildings) — BGS/UKHSA Indicative Atlas of Radon (Great Britain) and BGS/GSNI atlas (Northern Ireland). Six-band classification from "< 1%" to "> 30%" probability.

Architect and historical-figure attributions

61,527 buildings (12%) have at least one named architect, designer, builder, or historical figure extracted from the register description by a language model. Names have been canonicalised so common variants (e.g. "Sir Christopher Wren" / "Christopher Wren") are merged. 30,250 distinct names are captured in total.

The most-frequent attributions in the dataset:

• Sir Giles Gilbert Scott — 2,160 buildings
• Sir Edwin Lutyens — 403 buildings
• William Burn — 290 buildings
• Sir George Gilbert Scott — 282 buildings
• William Butterfield — 279 buildings
• Thomas Telford — 270 buildings
• G.E. Street — 270 buildings
• John Carr — 248 buildings

Browse the full set of named architects by following the people links from any building page.

How to explore

• Search buildings — by name, attributes, location, grade, type, period, style, material, or architect.
• Search applications — by LPA, decision status, date, or extracted physical-attribute tag.
• Stats & Indices — visualised national-level breakdowns of the dataset.
• Or follow links into one of the country indexes above to browse by Local Planning Authority.

Licence and attribution

Source data is reused under the licences of its original publishers:

• Listed building registers — Crown Copyright / Open Government Licence v3.0 (Historic England, HES, Cadw) and Crown Copyright (NIEA).
• Planning data — sourced via PlanIt.org.uk, originating from individual UK Local Planning Authority portals.
• EPC Register — Open Government Licence v3.0 (DLUHC and statistics.gov.scot).
• HM Land Registry Price Paid Data — Open Government Licence v3.0.
• Flood and radon data — Open Government Licence v3.0 / British Geological Survey licence terms as applicable.
• ONS Postcode Directory — Open Government Licence v3.0.

Aggregated, derived, and structured outputs published on this site (canonical attribute tags, address matching, person-name canonicalisation, the rewritten descriptions, and all linkage between data sources) are available for non-commercial research use. For commercial use or bulk data access, please get in touch.

Coverage gaps and known issues

We are transparent about what's missing or imperfect:

• Northern Ireland is not yet covered for EPCs (no open data) or for flood risk (commercially licensed data only).
• Match rates on EPC and Price Paid records reflect that most certificates and sales are for non-listed properties — they're the right denominator only when you're looking at coverage of listed-building-relevant records.
• PlanIt's decision-feed is patchy: applications are captured reliably but decision dates and outcomes can be delayed or missing on individual records. We use multiple decision-state fields to compensate.
• Architect attributions come from LLM extraction and can occasionally pick up bibliographic mentions or witness names rather than architects themselves; the canonicalised top names are reliable.

Spotted incorrect data? Flag it — we use feedback to improve the matching pipelines.