What am I looking at?
This is a network graph or node-link diagram representing the users of the TNR system, a continuous learning system for inclusive education specialists, and the different types of links between them. It shows the time range from September 2011 to June 2015. There are three different kinds of user profiles: researchers and multipliers, who initially co-designed the system, and "others", i.e. users that joined after the design phase. Link strength and decay are very simplistic: each kind of relation (commenting, liking, inviting, etc.) has the same weight and is added to the cumulative link weight. After the decay time, it is simply subtracted. Although simplistic, one can derive some insights that would probably be more difficult to gain in static visualizations or by calculating network strengths and other properties. For example, researchers (blue squares) and multipliers (orange triangles) at this stage of the development of the community are still crucial to hold the network together. Although there are links between "others", they decay relatively fast.
How was it made?
This one is fairly straightforward. It uses the regular force layout of Mike Bostock's D3 and jQuery (at the time, this was more convenient than plain vanilla JavaScript). The node and link data is extracted from TNR's Drupal database. I used R with the RMySQL and RJSONIO libraries, but any other method making some queries and formatting the output to JSON will do.
But why?
I wanted to explore whether relatively simple visualizations like a network graph animated over time have the potential to yield meaningful insights for people not familiar with quantitative methods of network analysis. Also, I wanted to play around with D3 and R.
If I would do it again...
This really only was a throwaway prototype in order to play with some at the time more recent version of D3. Of course, the usual suspects apply: more controls about the animation and node/link selection, a legend, etc.