A couple of small datasets of the Facebook network data are available on the Internet. None of them are particularly large or complete, but they do give us a reasonable snapshot of part of Facebook's network. As the Facebook graph is a private data source, this partial view is probably the best that we can hope for.
We'll get the data from the
Stanford Large Network Dataset Collection (http://snap.stanford.edu/data/). This contains a number of network datasets, from Facebook and Twitter, to road networks and citation networks. To do this, we'll download the facebook.tar.gz file from http://snap.stanford.edu/data/egonets-Facebook.html. Once it's on your computer, you can extract it. When I put it into the folder with my source code, it created a directory named facebook.
The directory contains 10 sets of files. Each group is based on one primary vertex (user), and each contains five files. For vertex 0, these files would be as follows:
0.edges: This contains the vertices that the primary one links to.0.circles: This contains the groupings that the user has created for his or her friends.0.feat: This contains the features of the vertices that the user is adjacent to and ones that are listed in0.edges.0.featnames: This contains the names of the features described in0.featand0.egofeat. For Facebook, these values have been anonymized.
For these purposes, we'll just use the *.edges files.
Now let's turn our attention to the data in the files and what they represent.