I recently attempted to visualise the labrinthine UCL departmental structure in the form of a Force-Directed Graph:
It has been remarked (not unfairly) that this is the sort of pattern that I might wear, were it printed on a shirt. If there are any screen printers, fabric designers or shirtmakers wanting to collaborate, I’ll happily sport UCL in abstract form on my torso and arms. Just say the word.
(Although there are no labels above, the colours roughly correspond to University/School/Faculty/Department/Subdepartment, although there are grey areas where a series of departments have a common umbrella, for example; when it gets to this fine a level, some of the colour decisions are slightly arbitrary).
On a slightly unrelated topic, while at the Edinburgh Festival, I picked up this rather nice greetings card:
It’s by Edinburgh designer Sharon Pringle; the pattern is called “Anster” and it’s part of her “Elie” collection. Now, I doubt she created her pattern using a Force-Directed Graph of a university organisational structure programmed in Processing – but it looks a lot like the network visualisations I see – and it looks neat. I am tempted to buy myself one of her mugs. (She also has the aforementioned greetings cards, and handsome cushions)
Hi there Martin. So glad you like your card and my mug design! The design is based on a small sketch I found in my husband’s notebook. He’s an architect and he was working out pathways between groups of buildings. I took his diagram, tweeked it a bit and repeated it and ended up with the finished design.
Hmmm… do you know what technique he used to create them? Or was it freehand? I’m based in an architecture school, and there are various techniques that people here use to understand space and how people interact with it…
Is this a flat network, or have you enforced any kind of hierarchy? Perhaps integrating a hierachical approach would be an interesting way of working with nested departments, research groups etc?
Hi Dan – the hierarchy is implicit in the links; in fact, that’s how I’ve defined them. A link exists between an entity and its immediate parent entity, and that is all. It might be interesting to introduce some sort of collaborative linking to designate real research links rather than just organisational structure. It would be cool to introduce people and their affiliations, research students, etc but that would get rather complex.
What was the data source for this, or did you do it by hand? Also, what was the ‘force’ on the links or was that just for layout purposes?
Might be interesting to take a data-mining approach and slurp each group’s homepage, extract the keywords, and then see which departments are linked by common terms even though they are from different faculty groupings…
Hi Jon. It’s taken from the UCL website, done by hand. I used an older version of this:
So I could probably update it and be a bit cleverer about it (explicitly assign each entity a hierarchical level rather than define that by link depth, as I have done).
It’s a force graph layout; the links are of equal strength if I recall.
I like your idea. Let’s do it! Then put it on a fabric print.
EDIT: This was the source: