Social Computing has brought Human Sciences and the Information Technology ever closer together. For instance, Sociology is concerned with, the connections between individuals, between social institutions, and between individuals and social institutions as well as their influence on each other (Giddens & Sutton, 2009). Pierre Bourdieu, the French Sociologist proposed the concept of “social capital” which means the benefit an individual gains thanks to their social network. In the recent years, because of the emergence of “virtual” social networks such as Facebook, and the interest of Computer Scientists in studying these networks, we are able to visually see and analyse human social networks, and to re-examine the Sociological concepts which concern human social networks and social capital. In this essay I analyse a small social network.
To produce this graph I used the Netvizz app on the Facebook platform in order to generate a Gephi file that has information on Friendships in a Facebook network. The file was then opened in the [Gephi app] (http://gephi.github.io) on the Windows OS to generate a graph. The layout was set to Forced Atlas 2; and then, Modularity Class was applied in order to give unique colours to different clusters. Labels were applied, and the degree range was limited to reduce clutter in the graph.
This social network consists of two clusters.
- A yellow cluster is on the right and a purple cluster is on the left.
- Each cluster is made up of 10 to 20 nodes.
- Each node represents an individual.
- These clusters are communities of individuals.
These communities can be geographical communities or other types such as communities of interest. The nodes are connected to each other with edges. According to Hansen, Shneiderman, and Smith (2010, p. 34) edges “are the building blocks of networks”.
- In this image, edges represent Facebook Friends.
- Edges are also known as ties.
- Ties can be strong or weak.
The ties in the yellow cluster are weaker than the purple cluster. This is because in the yellow cluster the nodes are connected to each other with fewer edges and not every node is connected to every other node in that cluster. In the purple cluster, on the other hand, more nodes are interconnected and this is evident by the higher number of edges that connect the nodes of the purple cluster. There is a very weak tie between the yellow cluster and the purple cluster. They are connected with only two edges.
Facebook is a Multimodal network since ties can be based on Likes, Comments, Shares, or just Friendships. However, for the purposes of this report, only Friendships are illustrated. The illustrated edges are undirected edges because these are mutual relationships [ibid]. Furthermore, the illustrated edges are unweighted edges because they are solely based on Friendship rather than the quality and strength of the Friendship.
Some of the nodes of the purple cluster have a high level of degree centrality. For instance, Caoimhe McCreanor at the bottom of the purple cluster has a high number of edges. On the other hand none of the nodes of the yellow cluster have a high level of degree centrality.
Betweenness centrality is a metric that measures to what degree removing a node would disconnect other nodes (Hansen et al., 2010, p. 40). In our graph Ger Corcoran has the highest level of betweenness centrality as he is a bridge between two clusters and removing him would “disrupt the connections between other people in the network” [ibid]. Ger Corcoran is an important node because he is a bridge between the two clusters. Shauna Doyle (below Ger) is also an important node as she is a bridge between the two subgroups in the purple cluster. Also, if Shauna is removed, many nodes at the bottom of the purple cluster will lose their connection with the yellow cluster.
The degree of clustering coefficient is much higher in the purple cluster where almost every node is connected to every other node. Clustering coefficient is lower in the yellow cluster where each node is tied to fewer nodes and so there is less concentration of edges in the yellow cluster.
With regards to eigenvector centrality, Ger Corcoran (purple), Joanne Hynes (yellow), and Sara Lowry (yellow) have a high value in the network. They are in a strategic position because being friends with them allows other nodes to be connected with many other nodes.
Giddens, A., & Sutton, P. W. (2009). Sociology (6th ed.). Cambridge: Polity. Hansen, D., Shneiderman, B., & Smith, M. A. (2010). Analyzing Social Media Networks with NodeXL : Insights from a Connected World. Burlington: Morgan Kaufmann.