Skip to content

Open & Linked Data for Universities

I’m currently at IWMW2010 which is going pretty well. I ran a 90 minute workshop yesterday on university linked-data for university web managers. My basic points were:

  • Get your URIs right. eg.
  • Start with DC, SKOS, SIOC, FOAF, GEO (I wish I had that list 5 years ago!)
  • Pick the easy stuff and do that first
  • Don’t focus on accurate modeling in the ontology — rather think about how people might use the data for something useful
  • CSV is much better than no raw data
  • RDFa is not the place to start learning RDF.
  • A linked-data manager does not need to understand the fine details RDF, any more than a web manager needs to understand HTML & CSS
  • Build data for your own consumption
  • You’re already paying someone somewhere to keep much of this data up to date, but they are just failing to share it in a useful way. Turning it into open data should save lots of people around the university recreating existing datasets.
  • Use a tool to check your RDF says what you meant
  • Don’t worry about OWL & ontologies. You are better off (initially) writing your ontology for humans to understand, rather than machines.
  • I am not my homepage! — as RDF uptake increases there will be more people confusing URLs of documents with URIs for things. We’ll cope with that from the great unwashed linked data sources, but there’s no reason to do it when you know better.

For my talk I did something which I wish could become standard practice. Make a web page containing all the links in the slides and give the audience a tiny URL to write down to save them trying scribble every URL down. I also owe thanks to Dr Nick Gibbins, who provided me his Intro to Linked Data slides and saved me a hell of a lot of work.

I’ve spent the past few months working on a brainstorm of all the possible datasets a university could consider publishing.

Over the session, and talking to people at IWMW, I’ve added a few new entries such as Reading Lists and data for accountability. Suggestions from the peanut gallery are positively encouraged!

I’ve also been busy converting Southampton’s list of “Common Learning Spaces” into RDF, via screen scraping and a whole bunch of RegExp’s. Next I plan to build a dataset of all the university buildings including latitude and longitude. It turns out this is quite easy to do, all I had to do was create a Google Map of University Buildings based on our campus map and then export it to Google Earth, which really produces a KML file which I can easily munge into RDF.

Given that data I should then be able to create a tool which mashes up the two sets of data and can produce a map of all teaching rooms on campus with movable seating, or a smart screen. Hopefully this will be a killer example of how very simple open & linked data can be a win for a university.

Posted in Best Practice, RDF.

2 Responses

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.

  1. Chris Rusbridge says

    Chris, ou may have seen from the tweetstream that there is some consternation at the complete lack of research data from you mind map. Even just one tiny leaf node would help remind us that research data exist! Trying to expand it might not be worth it in this context, as it might rapidly get terminally complex…

Continuing the Discussion

  1. IWMW10: Linked Data » Overdue Ideas linked to this post on July 14, 2010

    [...] Chris has helpfully published a page with all the relevant links from his talk, except the mindmap of institutional data that might be published. Chris has also helpfully blogged the Linked Data session. [...]

Some HTML is OK

or, reply to this post via trackback.