I saw this post on FlowingData and thought, "Hey, I can do that, let's give this Gephi thing a go."
I don't have that many Facebook friends, as I try to keep my network well maintained, and I'm also not a heavy user of the service. Also I've always kind of wondered - if you are one of those people who has 2000 Facebook friends, are they really all your 'friends'? If I put you (you silly teenage girl) in a room with those 2000 people, would you be able to call all of them by name? Remember where you met them? What their favorite color is? I digress.
The steps to producing the graph are simple:
1. Install the netvizz Facebook application
2. Run it
3. Import gdf file into Gephi
4. Wow! A graph!
As I said, I don't have that many Facebook friends but I still found the results pretty interesting:
Red is immediate family, green my Mom's side and blue my Dad's. The orange are university friends, and purple High School. Yellow are randoms and friends of friends. Teal is friends of my Mom's relatives, and pink friends of one of the immediate family. Light blue is a group of friends made while travelling.
The nodes are sized by degree.
Interesting point to note:
High school friend (purple, outlying from others) and friend of immediate family (pink, bottom right node) are both connected to friend of Mom's family (teal node, bottom) through events totally unrelated to the rest of the network. Small world.
This is that case when you add a new friend on Facebook and it says you already have a mutual friend, and you stop and think, 'Wait, we do? Sarah knows Thomas? But how did.... through who... when did.....? Huh.'