National Buildings Database: Non-Domestic Building Synthetic Population, 2023

Smith, Olly and Evans, Steve and Shamsi, Haris and Fennell, Pamela and Liddiard, Rob and Ruyssevelt, Paul and Neto-Bradley, André (2026). National Buildings Database: Non-Domestic Building Synthetic Population, 2023. [Data Collection]. Colchester, Essex: UK Data Service. 10.5255/UKDA-SN-858479

This research project combines big data analytics with targeted surveys to produce a digital twin database of the building stock. This live database represents a major step change in the department’s evidence base and will underpin the future National Buildings Model that provides evidence for policy making and analysis of energy use in buildings within the department. This new digital twin database offers added value for money with automated updating capability using existing data sources, supplanting the need to routinely repeat expensive surveys.

Data description (abstract)

This dataset contains a synthetic population of Great Britain’s non-domestic building stock in 2023 generated from the National Buildings Database (NBD). The NBD has been developed for DESNZ linking administrative and geospatial sources to premise-level attributes including sensitive annualised energy indicators derived from metering and related records. This has produced a database with a one-to-one representation of every premises and building in Great Britain, but owing to presence of sensitive and personal data in these records the microdata cannot be shared outside Government. This synthetic population provides a premises level synthetic and non-disclosive dataset of approximately 2.2 million non-domestic premises, which is designed for query and aggregation by both geography and non-domestic activity classification. The records include annualised energy indicators alongside core building attributes describing size, age and construction characteristics. The dataset is designed to preserve key marginal distributions, cross-attribute dependencies and spatial patterns while reducing disclosure risk.

Data creators:

Creator Name	Affiliation	ORCID (as URL)
Smith Olly	UCL
Evans Steve	UCL
Shamsi Haris	UCL
Fennell Pamela	UCL
Liddiard Rob	UCL
Ruyssevelt Paul	UCL
Neto-Bradley André	Department for Energy Security and Net Zero

Sponsors:

Department for Energy Security and Net Zero

Grant reference:

Contract Reference 2023/S 000-015901

Topic classification:

Housing and land use
Science and technology

Keywords:

BUILDINGS, ENERGY CONSUMPTION, ENERGY EFFICIENCY, SCIENCE AND TECHNOLOGY, HOUSING

Project title:

National Buildings Database

Project dates:

From	To
October 2023	January 2026

Date published:

29 Jun 2026 11:10

Last modified:

30 Jun 2026 11:35

Coverage and Methodology

Access and Administration

Data sourcing, processing and preparation:

This dataset is an anonymised synthetic derivative of DESNZ’s National Buildings Database (NBD), which combines Ordnance Survey address and mapping data, Valuation Office Agency non-domestic property records, EPC/DEC data, and linked non-domestic gas and electricity meter data. The anonymised dataset was derived from NBD using a diffusion-based synthesis process designed to preserve broad geographic and building-activity relationships while ensuring anonymity; where an activity-geographical combination contained fewer than 10 source records, the process fell back to broader groupings. Post-processing quality checks were also applied to check for near duplicates, with any failing combinations omitted to ensure non-disclosure.

Rights owners: