As my second Coursera MOOC, I’ve recently started Social Network Analysis. I don’t have enough time this spring to follow it all through, but anyway. One of the first assignments was to look at the Facebook network. Because I have signed the Coursera Honor Code of not to reveal anything about how I actually did the homework, here is something else instead. It is about anonymizing.
I had this idea of constructing a new attribute column in Gephi by regex-ing initials of friends, and using values in this column as node labels. During a Gephi session you can do this, but when you export the data as a GEXF file, the original labels surface again.
So, I made a brute force anonymizing by writing a small XSLT stylesheet, and running it against the GEXF file. To render the result on a web page, sigma.js seemed a nice tool to test.
