I have continued exploring ways in which one can visualize Facebook networks.  Here I have used the Netvizz application to extract networks of people who have joined a particular Facebook group.  You can get access to group network data for any group that you are a member of.  I decided to examine the ‘UCT – University of Cape Town’ group of which I am a member. At the time of extraction the group contained 2248 people.  People in the group are not friends by default, but there were many instances where people were friends, in total 5960 relationships existed in the network beyond people having joined the group.  When you first load the network data into Gephi is looks something like this.

You have to run some statistics and some algroithms to make things interesting.  I ran the Modularity Report which I believe attempts to identify communities within the network.  Once you run the report you can then partition the nodes by colour according to the communities discovered.  I then ran the Radial Axis Layout and grouped the nodes by modularity to generate this view.

The network turned out to be quite diverse with many disparate and small groups of friends represented in this group.  The clusters of groups of people are stacked around the circle and the lines between connect individuals within different groups.

We then ran the OpenOrd layout on the data and came up with this interesting visual.

We don’t know exactly what this particular layout is showing, but it is very interesting the way in which the nodes swell and settle when you run the program. Will continue exploring.


CC BY 4.0 Visualizing Facebook Groups by Michael Paskevicius is licensed under a Creative Commons Attribution 4.0 International License.