Skip to main content

Data sources at a glance

Each source has its own notes below: what it is, how we link it, and what you can and cannot infer from it. Glossary defines the terms. The signal column says how directly each source shows a working relationship.

Research collaborations

Provider: OpenAlex. Refreshed: monthly. Signal: strong.

What it is

OpenAlex is an open index of research papers and the organisations their authors belong to. A company is a research collaborator when at least one paper from the last 30 years lists authors from both the university and the company.
  • Only organisations that OpenAlex records as companies based in the UK are included.
  • We match each company to its Companies House record by name, including the other names OpenAlex knows it by.

What you can infer

  • The university and the company produced published research together. This is an R&D relationship, not a casual contact.
  • How many companies a university co-publishes with, and in which sectors.

What you can’t infer

  • Collaboration that produced no paper, such as consultancy, contract research or confidential work.
  • Collaboration with companies outside the UK.
  • How recent or how deep the relationship is. One paper in 30 years counts the same as many papers last year.
  • That every match is right. Name matching misses some companies and links a few to the wrong Companies House record.
  • Papers filed under a university’s former name in OpenAlex. For example, City St George’s, University of London counts only papers filed under its own name since its 2024 merger, not those of City or St George’s before it.

Innovate UK funding and KTPs

Provider: Innovate UK, part of UK Research and Innovation (UKRI). Refreshed: monthly. Signal: strong when present. See also Innovate UK data.

What it is

Innovate UK publishes every project it funds, with each organisation that took part. A company is an Innovate UK funding partner when it took part in the same funded project as the university. A Knowledge Transfer Partnership (KTP) is an Innovate UK programme in which a graduate works inside a company on a project the university supervises. KTP links are a subset of Innovate UK funding partnerships.
  • We read every Innovate UK project the university took part in, whatever the funding programme.
  • We link a partner through the company number on its own project record. Where the record gives none, we match the partner’s name to Companies House.
  • We read a KTP’s company from the project title, because KTP records often list only the university.
  • We leave out business-support programmes, such as the Enterprise Europe Network, trade missions and SME coaching. In these the university delivers support to businesses and does not partner with them.

What you can infer

  • The university and the company were funded to work on the same innovation project. This is strong evidence of a working relationship.
  • A KTP link shows the company hosted a graduate on a project the university supervised.

What you can’t infer

  • Partnerships funded elsewhere, such as research council grants.
  • That coverage is complete. A partner with no company number and an unclear name can be missed, as can a KTP whose title does not name its company. A few name matches are wrong.
  • The size of the grant or the company’s share of it.
  • That every KTP counts as one. Innovate UK files some KTPs under another programme, so they count as Innovate UK funding partnerships but not as KTPs.

Spinouts

Provider: Higher Education Statistics Agency (HESA) spin-out register, plus The Data City’s own checks. Refreshed: monthly. Signal: strong. See also University spinouts.

What it is

A spinout is a company set up to put a university’s research or knowledge to commercial use, and recognised as such. Most come from the HESA spin-out register, which universities report to. The rest come from The Data City’s own checks of university news, companies owned by university technology-transfer offices, and company websites. We assess each one against HESA’s definition of a spinout, and keep its evidence, before we add it.
  • We link each spinout to its university by the UK Provider Reference Number (UKPRN) on the register, or by the university’s name when we do not hold that UKPRN.

What you can infer

  • The company exists because of the university’s research or knowledge, as the university recognises.
  • When spinouts were founded, as the spinouts-by-year chart shows.

What you can’t infer

  • Staff or graduate start-ups that were not formal spinouts. Graduates who founded companies appear under alumni founders instead.
  • Investment raised, or the university’s shareholding.
  • Spinouts that a university has not reported to HESA and our checks have not found.

Alumni directors and founders

Provider: Lightcast profiles, Wikidata and Companies House. Refreshed: monthly. Signal: medium. See also Lightcast data.

What it is

Graduate profiles from Lightcast, built mostly from public LinkedIn profiles, and notable graduates recorded in Wikidata. We match them to the officers that Companies House lists for each company. Alumni founders are a subset of alumni directors.
  • Director, from Lightcast: a graduate whose current employer has a Companies House officer with the same surname, or whose current job title is senior.
  • Director, from Wikidata: a living notable graduate linked to a company in Wikidata who matches one of its Companies House officers, current or past, on surname and forename.
  • Founder: a director whose matched Companies House officer is flagged as a likely founder of the company, or whom Wikidata records as its founder.

What you can infer

  • Where a university’s graduates hold leadership roles, and which companies they helped found.

What you can’t infer

  • When they took the role or graduated. We use no dates for roles or study, so a director may have graduated decades ago.
  • That a role is current. A Wikidata match can be an officer who has since resigned.
  • That a match is verified. A common surname can match the wrong officer, and a senior job title counts even without a Companies House match. The employer’s match to its Companies House record can also be wrong.
  • Graduates with no public profile, or outside Lightcast’s data.
  • That every founder is counted. Not every founder is flagged as one.

Co-filed patents

Provider: European Patent Office. Refreshed: monthly. Signal: strong.

What it is

A patent family is one invention, however many patent offices it was filed in. A company co-filed a patent with the university when both are named as applicants on at least one family first filed since 2000, in the European Patent Office’s DOCDB bibliographic data.
  • We read British (GB), European (EP) and international (WO) filings.
  • We match each company applicant to Companies House. We leave out applicants based outside the UK, individuals and organisations we cannot match.

What you can infer

  • The university and the company jointly applied for an invention. This is a legal act, not a record of contact.
  • How many patent families the university has filed since 2000, with or without a company partner. The Patents panel and the leaderboard in UIIE show this portfolio.

What you can’t infer

  • A university’s portfolio from its partner count. Many families name the university alone, so they count in the portfolio but add no partner.
  • A total across universities. A family co-filed by two universities counts under each.
  • Per-office counts such as those in the Higher Education Business and Community Interaction (HE-BCI) survey. Counting families reads lower, because one family covers every office it was filed in.
  • Licences or transfers of a patent after it was filed.
  • Inventions first filed in about the last 18 months. Patent offices publish an application around 18 months after it is first filed, so the newest filings are not in the data yet.
This product contains data sourced from EPO databases, © European Patent Organisation.

Research council grants

Provider: UKRI Gateway to Research. Refreshed: monthly, from Gateway to Research’s monthly download. Signal: varies. It is strong for small consortia, and weak for training grants and large consortia.

What it is

Gateway to Research is UKRI’s public record of the grants its research councils fund, such as Engineering and Physical Sciences Research Council (EPSRC) and Biotechnology and Biological Sciences Research Council (BBSRC) grants. A company has a research council grant link when the university and the company are both named on the same grant, as the lead organisation or a project partner. A small-consortium grant is one with 10 or fewer project partners, not counting the lead organisation, that is not a training grant.
  • We read only research council grants. Innovate UK projects come from the Innovate UK data, as Innovate UK funding partnerships.
  • Most grant partners carry no company number, so we match each one to Companies House by name. A close name match, or a pick among several companies with the same name, counts only when something confirms it, such as the partner’s postcode matching an address the company uses. We drop a company incorporated after the grant ended, because it can only be a namesake.

What you can infer

  • The university and the company were named together on a publicly funded research grant.
  • From a small-consortium grant, a working relationship on a defined project. Studentships, which name one company, are in this group.

What you can’t infer

  • A working relationship from a training grant or a large consortium. A Centre for Doctoral Training or a consortium can name dozens of partners. Use the Small consortia only filter for the stronger evidence.
  • The value of a grant or the company’s share of it.
  • Collaboration that researchers report after a grant ends.
  • A total across universities. A grant that names two universities counts under each.
  • That every match is right. Name matching misses some partners, and a few same-name companies remain.
Source: UKRI Gateway to Research. Contains public sector information licensed under the Open Government Licence v3.0.

Graduate flows

Provider: Lightcast profiles and Companies House. Refreshed: monthly. Signal: indicative.

What it is

Lightcast graduate profiles, matched to the UK companies the graduates work at today. It is a snapshot of where people are now. It is not an official destinations survey, such as HESA Graduate Outcomes.
  • Each release covers graduates of the last five calendar years, counting the current one. The dashboard panel shows the exact years.
  • A person who studied at two universities is counted under each. Per-university totals therefore add up to more than the number of distinct people.

What you can infer

  • Which sectors and regions a university’s recent graduates work in.

What you can’t infer

  • The public sector (NHS, civil service), overseas employers or the self-employed. We do not count them.
  • Graduates outside the cohort window, or graduates with no usable graduation year.
  • A count of people from the sector chart. It counts employers. The region charts count people, and a graduate counts in every region where their employer has offices, so regional bars can total more than the headline.
  • That every university is covered. Some universities have no matched graduates yet.
Last modified on November 17, 2026