After learning about Social Network Analysis (SNA) in the Social Network Analysis MOOC of Lada Adamic (8 weeks in November and December 2014) I started to visualize my online groups with gephi. In today’s workshop I presented some key concepts of SNA and shared my experience of using them for my online groups with my colleagues.
After the introduction with a small story about my love of visualizations I started to define social networks, nodes and egdes and explained their relationship (directed and undirected edges, degrees and weights). I asked questions about nodes: are they well connected, what´s their function in the network? Or – are there any communities in the network? And I spoke about the Erdős-Rényi Graph and the Barabasi-Albert Model to simulate networks. I presented visualizations with netlogo explaining the appearance of the giant component in Erdős-Rényi Graphs and the apperance of hubs in the Barabasi-Albert Model.
The last part of theory included the explanations of degree centrality (1), betweenness centrality (2) and closeness (3) and a short overview how to detect a community.
After the coffee break we switched to „applied“ SNA. I explained how to prepare data for gephi and presented two data sets.
Dataset 1: My first approach to get data from students‘ interaction was via counting manually the interactivity of my students in our google+ community. I succeeded to count their activities during 4 days of November 2014 (and November wasn’t their most active month).
Dataset 2: In the next step I persuaded the administrator of our learning platform to automatically count how students communicate with each other in one of my courses. He provided me with two sets of data, one of the interaction during the first three weeks of their online socialization (October 2014) and one of the entire interaction during the winter semester 2014/15.
Visualization of the first dataset: Gephi created the network of the students in Google+, the size of the node shows the degree (how connected the person is). In gephi it is possible to click on one node and emphasize the interaction of this node with the other nodes.
Visualization of the second dataset – the phase of online socialization: In this visualization the size of the node ist connected with its degree, and the color with its betweenness.
In this graph I’m the blue node in the middle of the network. My node has the highest betweenness, it is in a „broker“ situation between two groups of students. Stud21 has a high closeness centrality.
During the phase of online socialization it is ok for the teacher to occupy such an important position. But during the semester the students should gain more independence, so let’s take a look at the social network analysis after the end of the semester.
Visualization of the second dataset – at the end of the semester: In this visualization the blue node has lost its broker position. The students started to interact intensively with each other, e.g. the green ones are now important players of the network as well.
At the end of the workshop we discussed my conclusions of the study of SNA and the application of SNA for my online groups:
- Using SNA I’m getting nice graphs and some insights – but there’s a lot I still don’t understand.
- SNA has big potential but is rather complex.
- I want to learn how to better use gephi.
- In future I would like to generate other data for visualization (e.g. twitter data).
Slides (jn German)