In this 2010 article, Kumar, Novak, and Tomkins examine the evolution of large online networks. The researchers look at longitudinal friendship networks on Flickr and Yahoo! 360, and examine how they change over time. Instead of taking snapshots of the network, they use digital trace data to create “timegraphs”, which represent the current state of the network at any given time.
They find that the density of the network (i.e., the ratio of edges to nodes) follows a similar – and unexpected – pattern for both networks. It increases very quickly initially, then decreases, then increases gradually. They believe this indicates something like the hype cycle, where there is a lot of initial excitement, followed by some disillusionment, and finally steady acceptance and growth.
They also examine the difference between users connected to the giant component, and those who are in other components. They find that those outside of the giant component form “stars”, with one user at the center, and others all connected to that user. They posit that this structure is the result of invitations – users invite other users to connect to these sites, and often the invitees will create an account, connect to the inviter, and not connect to anyone else. It’s a very good example of how network data can lead to the discovery of an otherwise hidden behavioral pattern.