Suffrago UK Data Hub
RankingsVerificationMethods

Independent service run by Suffrago – not an official government website. Every figure comes from an official UK publisher under its open licence and is checked against that source each time the data is refreshed.

Methods

How the data gets here, and how you can check it.

Official sources only

Every dataset comes straight from the publisher's own API or download: the Office for National Statistics, UK Parliament, government departments and agencies on GOV.UK, NHS England, the Department for Education, the Home Office's police.uk, the Office for Health Improvement and Disparities, the UK Health Security Agency, HM Land Registry and the National Energy System Operator. Nothing is typed in by hand and nothing comes from a third-party aggregator.

What is recorded for every dataset

  • The publisher and the public page where you can check the figures yourself.
  • The exact URL of every file fetched, its size and its SHA-256 fingerprint; the raw files are kept alongside the data.
  • The licence – the Open Government Licence v3.0, the Open Parliament Licence v3.0 or, for NESO, CC BY 4.0 – and the attribution it requires.
  • The date and time the data was retrieved, and the publisher's own release dates where it gives them.

Every figure in one file

Every figure on this site but those by small area is in one SQLite database you can download, zipped, with the part of the UK it covers, its evidence layer and its dataset's source, licence and retrieval time. Unzip it to open hub.sqlite. Its latest_figures view gives one row per figure and its datasets view one row per dataset. It is rebuilt with the site each time the data is refreshed and holds only the figures the site shows. Each dataset by small area has its own download on its page, its CSV and JSON zipped together.

How every figure is verified

  1. Provenance – publisher, source page, licence and retrieval time must all be present.
  2. Integrity – the stored raw files must still match their fingerprints, and re-reading them must reproduce every figure exactly.
  3. Live source – every source URL is downloaded again and every figure compared. Live counts that move by the minute, such as petition signatures, pass when the live value has not fallen and has risen by no more than 5%.
  4. Second official route – where the publisher offers the same numbers another way (the ONS CSV download as well as its JSON, a per-party member search as well as the party totals, HM Land Registry's Average price download as well as the SPARQL query), those are compared too. Two of the health figures from Fingertips are also compared with their original publishers: healthy life expectancy with the ONS's own workbook and Year 6 obesity with the National Child Measurement Programme's data tables.
  5. Source page – the page named as the source must answer.

All of this runs on GitHub Actions. Every morning it checks each dataset whose publisher has released new figures since, and each dataset on its own day of the week, so every figure is checked against its source at least once a week; one that fails is checked again every morning until it passes. When the hub's own code changes in between, every stored file is re-read with the new code before the change goes live, and any dataset it would show differently is downloaded and checked against its source again first.

Results are on the verification page and on each dataset page.

DWP benefit caseloads from Stat-Xplore

DWP publishes its benefit caseloads in full only on Stat-Xplore, which needs a key. Only one step of the daily refresh on GitHub Actions can use it. That step queries Stat-Xplore, queries it again and compares every figure, re-reads the response, and compares the latest figures with DWP's own release on GOV.UK. It hands on only the checked figures and their proof; Stat-Xplore's raw responses are not kept. DWP adds a small random amount to every Stat-Xplore figure to protect confidentiality, so a dataset's nations and regions may not add up exactly to its total. These pages appear only once a refresh has verified them.

Not an official service

This site is built by Suffrago to bring official statistics together in one place. It is not run by or on behalf of the UK Government, and the publishers named here do not endorse it. Where figures are revised by their publisher, this site shows them as they stood at the retrieval time given on each page.