Here's an excerpt from the JISC's Monitoring Unit dataset that lists all UK Universities and Colleges, their lat and long and postcode:

http://tinyurl.com/68smyh8

The response comes from a query against the SPARQL endpoint of an RDF version of the JISCmu dataset hosted at http://api.talis.com/stores/jisc-prod-dev1/services/sparql The query has been piped through the SPARQL client at http://data-gov.tw.rpi.edu/ws/sparqlproxy.php because it gives CSV as an output format.

The JISC MU dataset itself only holds a single geographical point per institution, but it is linked in the Talis store to freebase.com identifiers and excerpts, which may hold more in some cases. You can get at freebase.com data by sticking ?muOrg owl:sameAs ?fbID . in the query below

The SPARQL query itself is:

PREFIX mu:   <http://www.jiscmu.ac.uk/schema/muweb/>
PREFIX geo: <http://www.w3.org/2003/01/geo/wgs84_pos#>
SELECT ?name ?postcode ?lat ?long
WHERE {  
?muOrg a <http://www.jiscmu.ac.uk/schema/muweb/Organisation> .
{ ?muOrg mu:Sector "Further Education" } UNION { ?muOrg mu:Sector "Higher Education" } .
?muOrg mu:OfficialName ?name .
?muOrg mu:PostCode ?postcode .
?muOrg geo:lat ?lat .
?muOrg geo:long ?long .
}

The JISC Monitoring Unit (JISCMU) API may have what you need, at least for HE?

See also the Linked Data conversion that Jo Walsh/@Metazool put together: Institutional locations as Linked Data through Google Refine

Something I noticed looking at HESA coding schemes was that as well as institutional identifiers there are campus identifiers. Something I don't know offhand is whether the JISCMU data includes HESA institutional identifiers, and "central" postcodes for HEIs, or if it goes as far as providing postcode/location data for campuses along with HESA campus identifiers?

There are also UCAS identifier codes - I'm not sure where the official list is, but there is a mapping from university name to UCAS identifier on the Guardian Datastore Rosetta data spreadsheet.

There must be mappings from HESA to UCAS codes somewhere? Though maybe easy enough to do via reconciliation on institution name?

The UKRLP holds all of this data - including other sectors such as schools and training agencies:

http://www.ukrlp.co.uk/

However I think you'd need to scrape the data from the site as it doesn't have an API as far as I'm aware.

Follow up to my earlier response:

You can find out the shape of the JISC MU dataset by running something like:

DESCRIBE <http://data.jiscmu.ac.uk/rest/organisations/id/83>

Against the SPARQL endpoint at http://api.talis.com/stores/jisc-prod-dev1/services/sparql

All the data in that set are organisations with those kinds of properties. Yes, they do include both HESA and UCAS codes.

To get at something like the domain name of each organisation, try:

PREFIX mu:   <http://www.jiscmu.ac.uk/schema/muweb/>
SELECT DISTINCT ?name ?domainname
WHERE {  
?muOrg a <http://www.jiscmu.ac.uk/schema/muweb/Organisation> .
{ ?muOrg mu:Sector "Further Education" } UNION { ?muOrg mu:Sector "Higher Education" } .
?muOrg mu:OfficialName ?name .
?muOrg mu:domainName ?domainname .
}

Running the query via http://data-gov.tw.rpi.edu/ws/sparqlproxy.php can give you html and spreadsheets

Note: as of this "touch time" the links below return nothing. Thus, the remote store might be undergoing maintenance of the data has moved. Coincidentally I made a similar demo elsewhere [1] a few minutes ago.

Here are some variations of this example courtesy of Virtuoso and its support of SPARQL-FED:

1. http://uriburner.com/c/ELYKIM -- raw SPARQL Protocol URL that returns an HTML based results doc

2. http://uriburner.com/c/ENN4SN -- iSPARQL generated Linked Data Page

3. https://spreadsheets.google.com/ccc?key=0AihbIyhlsQSxdDBsczF0UjJoMGZrV21BbnlHZnczTEE&hl=en -- Google Spreadsheet that consumed SPARQL-FED URL leveraging spreadsheet output format for SPARQL query results.

If I have a URL for the entire data set, the demonstration can be expanded across many other dimensions that includes joins across the LOD cloud plus facets driven follow-your-nose oriented pages etc..

Links:

1. http://answers.semanticweb.com/questions/9173/different-results-at-differnt-endpoints-while-using-sparql-query?page=1#9181 -- A variation of above the only differs re. target data, the SPARQL-FED and SPARQL-Routing concepts still apply.

Now that I've loaded the data into our URIBurner [1] and LOD Cloud cache [2] instances of Virtuoso, here are some additional demo links:

1. http://uriburner.com/describe/?url=http://data.jiscmu.ac.uk/rest/organisations/id/320 -- description of an Entity via a Faceted Navigation and Exploration oriented page
2. http://lod.openlinksw.com/describe/?url=http://data.jiscmu.ac.uk/rest/organisations/id/320 -- ditto but using LOD cloud cache which has a massive Linked Data set that helps improve the data via aiding identification of other URI cross references
3. http://bit.ly/lMDmRx -- example showing how to discover URIs for postcodes from the Ordnance Survey dataset via results of a text pattern search.

Kingsley