Dana House · Open Data

Dana House Open Data

A harmonised, enriched dataset of active charities registered across England and Wales, Scotland, and Northern Ireland, annotated with cause classifications and financial indicators.

Cite this dataset
@dataset{danahouse2026,
  title     = {Dana House Open Data: UK Charity Register, Enriched},
  author    = {{Dana House}},
  year      = {2026},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.PENDING},
  url       = {https://zenodo.org/record/PENDING},
  version   = {v0-pre-release},
  license   = {CC-BY-4.0}
}

DOI 10.5281/zenodo.PENDING activates on first release. The Zenodo record will resolve at that point.

Dataset card

Description

This dataset harmonises the public charity registers maintained by the Charity Commission for England and Wales (CCEW), the Office of the Scottish Charity Regulator (OSCR), and the Charity Commission for Northern Ireland (CCNI). Records are linked across registries using entity resolution, augmented with website evidence where available, and annotated with cause classifications from a 111-cause taxonomy.

The dataset is intended for researchers studying civil society, third-sector economics, philanthropic allocation, and charitable organisational behaviour in the UK context.

Coverage

Population: Active charities across all three UK charity registers as of the dataset snapshot date. The v1 release covers approximately200,000+ active organisations.

Geographic scope: United Kingdom (England, Wales, Scotland, Northern Ireland).

Time range: Charities registered from circa 1850 to snapshot date. Registration dates are populated where available from source registers.

Schema

The v1 schema includes cause classifications. A Phase 1 diagnosis (2026-05-05) confirmed 95.6% cause coverage with zero records in review state. The classification_quality_band field allows researchers to filter by confidence level. See the Limitations section for guidance on appropriate use by band. 1

FieldTypeNullableDescription
charity_idintegernoInternal unique identifier, stable across releases.
charity_numbertextnoOfficial registration number as issued by the registering body.
nametextnoRegistered charity name at time of snapshot.
jurisdictiontextnoRegistering body. Values: ccew, oscr, ccni.
registration_datedateyesDate of registration with the relevant charity regulator.
removal_datedateyesDate of removal from the register. Null for currently active charities.
statustextyesCurrent registry status as reported by the source register.
income_bandtextyesFinancial size band: micro, small, medium, large, major. Null if income unavailable.
latest_incomenumericyesMost recent annual income in GBP. Null if not filed or not extractable.
latest_spendingnumericyesMost recent annual spending in GBP. Null if not filed or not extractable.
primary_causetextyesPrimary cause from the 111-cause taxonomy. Null for 4.4% of charities that could not be classified. See classification_quality_band.
classification_quality_bandtextyesConfidence band: high (15.4%), medium (49.9%), low (34.7%), null for unclassified. See Limitations before using low-band records.
websitetextyesRegistered or discovered website URL. Null for charities with no identified web presence.
legal_formtextyesLegal form as recorded by the regulator (e.g. charitable company, CIO, unincorporated association).
company_numbertextyesCompanies House number, where the charity is also registered as a company.
record_readiness_bandtextnoPipeline quality gate: premium, production, or production_limited.
License

This dataset is released under the Creative Commons Attribution 4.0 International License (SPDX: CC-BY-4.0).

You are free to share and adapt the material for any purpose, including commercial use, provided you give appropriate credit. Attribution should reference the dataset using the citation format above and include the DOI once it activates.

Derivative datasets must carry the same attribution notice. The source registers (CCEW, OSCR, CCNI) remain subject to their own Open Government Licence terms, which are compatible with CC BY 4.0.

Versioning

Current version: v0 (pre-release)

Each release receives a new version number and a new Zenodo DOI. The previous release remains accessible at its original DOI. Release cadence is not yet fixed while the dataset is pre-release, and each release will record what changed since the previous one.

Limitations and known issues
  1. Cause classifications are produced by a hybrid rule-based classifier without a formal hand-labelled ground truth set. Correctness is structurally assured for the high band (multi-source, high-margin signals) but unverified for medium and low bands. Low-band classifications should not be treated as authoritative.
  2. 4.4% of active charities (approximately 8,800 records) have no cause classification. These are organisations whose mission text is too generic, ambiguous, or absent for a defensible assignment.
  3. Income and spending figures are drawn from annual accounts filed with the regulator. Filing delays and voluntary filing thresholds mean many smaller charities have incomplete or absent financial data.
  4. Website evidence is available for approximately 63% of classified charities. Charities without website evidence are classified from registry text alone, which is typically shorter and less specific.
  5. Entity resolution across registries is incomplete. Some charities operating across multiple jurisdictions may appear as distinct records.
  6. The dataset reflects a point-in-time snapshot. Charities registered, removed, or materially changed after the snapshot date are not reflected.

Downloads

Dataset files will be made available here on first release. All files are mirrored to Zenodo at release and carry a DOI.

FileFormatSizeSHA-256Link
dana-house-open-data-v1.csv---Dataset downloads activate on first release.

Zenodo deposit

Each release of this dataset is deposited to Zenodo, a CERN-operated open-access repository, and assigned a permanent DOI. The Zenodo record will be linked here once the first release is deposited.

Using the DOI in citations ensures your reference remains stable and resolvable even if the dataset URL changes. The Zenodo record also carries archival metadata in DataCite schema, which is indexed by DataCite, OpenAIRE, and Google Dataset Search.

How it works

Full detail on how the dataset is constructed, validated, and released is available on the How it works page. That page documents the source stack, record matching approach, classification system, quality gates, and known blanks.

Contact for academic users

Researchers with questions about the dataset, or who would like to be notified of new releases, can email contact@danahouse.co.uk.

Notes

A full classification diagnosis is available in the project repository at docs/classification-diagnosis-2026-05-05.md. It covers population baseline, confidence distribution, per-cause coverage, pipeline run history, and the absence of a formal ground truth validation set.