Is there a better map of UK higher education providers than this 2005 one from the University of Wolverhampton?
That question prompted much debate on Twitter among HE data experts about what counts as a provider, where they are, and which dataset to use. Two people even created their own alternative maps on the spot. Some think we should crowdsource data into an open Google Doc that then feeds a map. Others don’t think we need a map at all.
What’s a higher education provider?
This is a surprisingly complex question and one that’s foxed many people over the years. The (in)famous Regulatory Partnership Group pretty-much gave up on the question in its 2014 Corporate Forms project, concluding that whatever entity it is that the student registers with it must have some kind of governing body that is legally accountable – language that carries forward into the Higher Education and Research Act and the Office for Students regulatory framework.
But are we talking about a single institution, its registered address, or all of its campuses? Is it a location, a physical thing such as a building, or an organisation? What about non-UK parent companies setting up providers over here? And what about college HE?
For the Universities and Colleges Admissions Service (UCAS), an institution is something you can apply to. But for the Higher Education Statistics Agency (HESA) an institution is something the funding council/regulator has defined as an institution. Yet even these two organisations define institution differently at different points. And in the National Student Survey, there is a distinction between place of tuition and place of registration. Things like joint medical schools are rarely legal entities but exist as separate institutions in UCAS, but not in HESA. UK Provider Reference Numbers (UKPRNs) define institutions as legal entities.
In data terms, HE in FE is a real mess. “Franchised” HE courses with FE colleges are found in HESA data but often listed as a separate institution by UCAS. If the students are actually registered as students of the college (sometimes referred to as “directly funded”) then all their data is collected by the Education and Skills Funding Agency (in England – or equivalent, if different collections in the rest of the UK). In fact, there is no equivalent in Wales as the directly-funded colleges are on HESA and subscribe to it. Much data is collected through the FE funding bodies using different definitions and colleges don’t have to pay in the way that HEIs and alternative providers have to subscribe to HESA. There’s quite a discrepancy in the subscriptions paid by the two sectors. The Office of the Independent Adjudicator deals with franchises in a particular way, and the rules for this are different to those of the Student Loans Company (SLC) for example. Franchise providers can’t be regulated in Wales – they only get student support through a university partner.
Student support regulations also vary according to the data used, with complexities around the types of institutions and how they are designated. Definitions can be different based on where a provider is located and the mode of study.
And where are they?
Once you’ve – sort of – decided what a provider is, how do you know where they are? By their postcode, or grid reference (Northing and Easting or longitude and latitude)? It can be useful for an institution to decide on a single reference point, such as the University of Southampton’s SO17 1BJ postcode. For a building, it could be the middle or the main entrance.
There’s a rumour that David Willetts’ driver’s satnav once took them to a sorting office on an industrial estate somewhere in the Midlands when he visited the University of Reading, ending up at the PO Box where the institution’s postcode is set. And if you take the average coordinates from institutions with multiple campuses, such as the University of Nottingham which has a presence in China and Malaysia, you’ll end up in the sea. Universities teach, research and examine in a huge range of locations, from local music venues to parks, to the Marianas Trench.
The formerly Jisc-funded project Data.ac.uk has learning provider data. As part of that, http://learning-provider.data.ac.uk is an unfunded project at the University of Southampton built with basic metadata from the UK Register of Learning Providers (UKRLP), linked to other datasets such as HESA, GatewayToResearch, DBPedia, and Unistats. It also features an unofficial dataset linking universities to their consortia. However, it only provides a postcode and the latitude/longitude of that postcode, rather than geographic information system (GIS) mapping data. HESA data on campuses lists Eastings and Northings for each campus of each institution, but not for alternative providers (yet). Provider information meta-data including coding frames, geoinformation and groups are all on the HESA Open Data Strategy but haven’t been released yet.
Which dataset is best?
In addition to Wolverhampton’s classic map, UCAS have one from last summer for applicants in PDF, @Dan_HE_man has a map of where he’s visited, and Push has this one from 2010. HESA’s Heidi Plus service can visualise institutions, but only using HESA data and privately to subscribers. Wales has a map of (some) providers here.
Possible datasets to use include:
- The Office for Students register (182 providers and counting)
- UK Register of Learning Providers (UKRLP)
- Unistats (for the full dataset, see the Unistats page on the HESA website)
- HESA campuses data lists Eastings and Northings (grid references) and postcodes for each campus of each institution (currently charged for)
- UCAS data with postcodes of the main campuses of all providers who are registered with UCAS for the 2019 cycle which you can apply to (not always legal entities)
- Quality Assurance Agency for Higher Education (QAA) data
- Each devolved government publishes a recognised and listed bodies list
- HEFCW list of directly-funded providers of higher education in Wales
- Scottish Funding Council institution campuses, via HESA data
- Universities Scotland members who provided their address details
- SLC’s database is private
- The Home Office’s Tier 4 register of licensed sponsors
It turns out not one of these sources is definitive, meaning an up-to-date map will remain a policy purist’s dream. And maybe that doesn’t matter, close enough is good enough for most uses.
HESA used to have a list of the institutions it collects data from (the INSTID field) but the list is so fluid now, especially since the inclusion of alternative providers, that the list of institutions changes while the data is being collected.
A better map
With special thanks to top data gurus Hamish McAlpine, Emma Rączka, Andy Youell, Dan Cook, David Best and Christopher Gutteridge, here’s an updated map using the latest data from eight of those datasets (each of which you can filter for). You can see all the underlying data and even copy it to have fun with your own map.
Unfortunately, it will quickly be out of date as OfS has yet to show any ambition of providing an API or similar real-time data stream (although I hear learning-provider.data.ac.uk might be freely available to a loving home).
With the sector’s regulators, funders and designated bodies currently going through a process of (hopefully creative) destruction, there are big decisions being made about data. The belated OfS data strategy could mark a significant milestone in this process and I hear that HESA’s data futures programme is racing ahead, though perhaps not taking everyone with it.
There’s a growing need for standard identifiers such as UKPRNs, and general naming consistency. The sector is (mostly) quite good at using identifiers, but the real world is messy, so maybe we just have to accept no definitive list of providers can exist. More data should also be open and editable, rather than locked away in PDFs.
And what about research data, could Research England map institutions’ funding, institutional engagement managers and grant tables in one place? And what about the potential to link HE datasets with other external sources to increase their usefulness e.g. the Smart Specialisation Hub’s map of England’s innovation system.
Another potential issue is data integration, could we produce a single source of research and teaching-related data? HESA doesn’t currently capture all research institutions. UK Research and Innovation (UKRI) has great potential to bring together data from the seven research councils, Innovate UK and Research England (and the OfS under its collaboration agreement with UKRI). But what should this look like?
So data’s important, but does a map even matter? Some use cases could be argued, for instance when looking at social mobility coldspots, or for travelling politicians/journalists/minstrels. Sometimes it’s nice to just be a bit creative about communicating multiple bewildering datasets in a more accessible manner, and who doesn’t like pretty pictures?
As a visualisation of a dataset, a map (and the underlying data) can only be useful for certain uses. Multiple datasets (and the maps they create) may be correct, but the needs of an anxious 17-year old are (probably) different to a 42-year old minister. Which in turn are different to a civil servant’s exacting requirements, or a commercial property firm’s interests, or a university regulator’s needs.
However, barring the minister, the powers that be certainly don’t seem to think maps matter. Neither funding bodies nor HESA ever did an official map, and don’t pin any hopes on OfS.
Hopefully, some of the above questions, rather than just the pretty poster, will be attracting the English minister’s eye.