Printer friendly version
Facebook is a community
25 January 2012
Researchers in Italy have used two high-speed computer algorithms to analyse the connections between a large sub-set of the more than half a billion users of the social networking site Facebook to reveal that the system has a very strong structure. The study, published in the International Journal of Social Network Mining, shows that Facebook has a well-defined community structure that follows a statistical power law in which there are a huge number of people with few connections and a much smaller number with a large number of connections.
Emilio Ferrara of the Department of Mathematics, at the University of Messina, has anonymised Facebook data and used two sophisticated algorithms to uncover the hidden network structure across Facebook's millions of users. His research demonstrates that as with many social networks in the everyday world and networks found in nature, Facebook has the three common properties of such systems. First, it demonstrates the "small world" effect, known colloquially as "six degrees of separation" in which it is frequently possible to connect the majority of members, the nodes, of a network with all the other members through a small number of mutual friends or connections.
Secondly, Facebook follows the power law degree distribution where there are many users with a small number of connections. There are thus fewer and fewer users with more and more connections and only a very small number of people with a huge number of connections. Thirdly, Facebook rather obviously manifests as a community of interacting users rather than a collection of individuals.
One might imagine that so much is obvious given the popularity and activity of Facebook, which is the number one web destination and "application" for many millions of people. However, in order to prove that it is indeed a community-type network a statistical analysis of the type carried out by Ferrara was required. With the proof in hand, one might now investigate the structure of the Facebook network in more detail, apply the findings to other social networks, such as Twitter and LinkedIn in order to spot the differences and similarities with a view to informing those who operate and create such networks. The same research might also point the way to a better understanding of natural networks, such as offline human communities, insect colonies or even the spread of emergent diseases.