How these figures are calculated

Source

The Energy Performance of Buildings Register, published by the Ministry of Housing, Communities and Local Government. It holds every domestic Energy Performance Certificate lodged in England and Wales.

Register content is Crown copyright and database right, made available under the Open Government Licence v3.0. Address components carry Royal Mail copyright and are shown live from the register at the moment you search; they are not stored here.

Certificates in our area

1,011,923

Postcode districts covered

121

Data last refreshed

8 August 2026

How a district figure is produced

1. The full national register is downloaded and filtered to the B, DY, WS and WV postcode areas.
2. Duplicates are removed. A property assessed several times over the years appears several times in the raw data. Only the most recent certificate per property is kept, matched on UPRN where present. Roughly 30% of raw rows are removed at this step. Skipping it would overstate the stock substantially, because improved properties are reassessed more often than neglected ones.
3. Properties are grouped by postcode district and the counts, medians and percentages are computed directly from the certificates in that group.
4. Districts with fewer than 250 certificates are excluded, as the statistics would be noise.

Currently 551,477 properties sit below band C across the area, and 55,152 cannot reach it even if every measure recommended on their certificate is carried out.

Where the improvement figures come from

On an individual property page, the route to band C is taken directly from that property’s own certificate. The register records the resulting rating after each recommended step, so the figures shown are the assessor’s, not ours.

Those steps are cumulative and order-dependent. Step two’s gain assumes step one has been done. They cannot be added together in arbitrary combinations, which is why this site presents them as a sequence rather than a menu of options.

The typical gains shown on district pages are different, and weaker. They are estimated across 23 measures by non-negative least squares over certificates lodged in the last five years, using the fact that a property’s potential score assumes all its recommendations applied. The model explains around three quarters of the variation, so the figures are shown as ranges and described as typical. They are a guide to what a measure is usually worth in this stock — not a prediction for any one property.

What these figures are not

• Not a complete picture of the housing stock. Only properties with a lodged certificate appear. A home never sold or let since 2008 has no EPC and is absent entirely.
• Not necessarily current. A certificate reflects the property as assessed on that date. Work carried out since will not show until it is reassessed.
• Not a survey. An EPC assesses energy performance using the standard RdSAP methodology. It is not a condition report, a valuation, or advice on whether a particular measure is right for a particular building.
• Not authoritative for compliance. The open data service is published for analysis. For a definitive position on any property, the register itself is the source, and a lodged certificate requires a physical survey.
• Averages are flattered by reassessment. Because only the latest certificate is kept, improved properties are counted at their improved score. The true condition of the stock is likely somewhat worse than these figures suggest.

Checking and corrections

Every published figure is recomputed from the source data and compared against what the page renders before each release, so a number on this site should always be reproducible from the register.

If you believe something here is wrong, please say so and quote the page. Errors get corrected and the correction gets noted — that is more useful to everyone than being right by default.