In a previous post (Visualising OU Courses and Qualifications), I showed how IBM's Many Eyes data visualisation service could be used to visualise OU course data, in particular the network of relationships between courses of qualifications.
There's a meeting at the OU on Monday where one of the agenda items is to look at a "single" curriculum for our ICT and Computing offerings, so I added a few more qualifications to the data list in case anyone wanted to make use of the visualisation to aid the meeting. (I didn't add all the Computing degrees - they seem to have down down the line of spawning "Computing with X" courses which will just clutter the diagram. I leave the visualisation of those qualifications as an exercise for members of the Computing department ;-)
The new data set can be found here : OU ICT C&Q data. Note that is is not authoritative and study choices should not be based on it...
Here are a couple of visualisations based on the data:
- Network visualisation
- Treemap visualisation
The network map is a little bit more useful with the additional data - node size now starts to reflects the degree of connectivity of elements:
It would be handy to be able to associate an additional text column with entries in each data column containing tooltip text for each item (even better if the text data could be pulled in from another, separate table; I've just got a freebase account and my first impression is that it's going to be like dabble db on steroids when it comes to cross-referencing and pivoting about data).
The Many Eyes treemap visualisation go the C&Q data is also starting to demonstrate it's power a little more obviously now, as the size of the data set increases: Treemap visualisation.
For example, here is a view of courses associated with qualifications, ordered by qualification:
This display makes it easy to see what courses are associated with any particular qualification.
It's also trivial to reorder the treemap to view of the qualifications associated with a particular course:
This view of the data makes it obvious which courses are key to a large number of programmes - and hence might also help identify ways of reducing the number of overlapping qualification offerings.
To swap between the two views, just drag and rearrange the data labels in the bar at the top of the screen:
When in a particular view, the search tool allows you to see the distribution of child elements across the parent elements. So for example, in the qualifications view, we can search on a course code to identify which qualifications admit that course as an option:
This search highlighting essentially adds another dimension to the visualisation, and one that I think may be useful in a visual course catalogue browser.
It would be even more powerful if it allowed Boolean search terms (this OR that).
Speaking of which - getting authoritative data about our courses and quals is decidedly non-trivial. It seems that different units within the OU have "ownership" over different parts of the data, with the result that people are unwilling to share their data. I also suspect that there are multiple copies of various bits of data (nominally the same) which are separately maintained.... Oops...
There is (at least one!) XCRI project underway at the OU, so it would be nice to think it can help establish an authoritative (and public!) view over the course catalogue for people to innovate with.
To drive innovation, I guess we'd really need to set up an affiliate scheme, too, so that anyone building a course choice catalogue on top of our data can get a referral fee if a student uses it to register on a course. (Maybe the referral fee would be held in escrow until the student completed the course...?)
Posted by ajh59 at July 28, 2007 12:22 PM