| FULLDOME | Facebook Visualization
This semester I finally took the chance to play with the Powerdome at the FH-Potsdam. It’s a classic 360° planetarium with some high-tech to display live generated content. Perfect for interactive systems. So I joined the class “Immersive Datenvisualisierung” guided by Boris Müller in which we had to pick a dataset and beam it onto the powerdome walls.
Inspired by the wolfram alpha facebook report which just came out to that time, I decided to play with facebook data. More precisely: Visualize friend networks.
Summed up roughly, I extracted my own facebook friendlist via the giveMeMyData in a machine readable JSON format and visualized it via processing. The system I used here is based on a force directional graph. With the help of toxiclibs (thanks to Daniel Shiffman for the great tutorials on this one! Go buy his book “The Nature of Code”!) I created a network node for every friend in my list with a negative force field, so they will always try to seperate. The second rule applied is: If two friends of mine are mutual friends as well, they will attract each other. These two basic rules form clusters of people that know each other.
When I first saw the system working, I was pretty blown away how much these clusters that began to form, visualize exactly the mental model of groups of firends I had in my head.
Next I added a bit more information. I read out the affiliations. These are institutions, mainly schools, that people have been to. The more people went to a place the bigger it is. The place’s position in the network is averaged between the peoples positions. And as expected, the friend clusters have a strong connection to the institutions. Also I realized that these places work like cluster-labels. They make it much easier to connect to this otherwise anonymus friend network because all of the sudden it’s put in context.
Did I mention the whole network is renderend as a three dimensional system!? Thanks to Christopher Warnow, it is even distorted correctly, so the illusion is perfect. It feels like standing in the middle of a giant star-field like network system!
But how to navigate this? Remember, we are in 360° context. There is no left and right, front or back. In the end I decided to use the iPhone and it’s compass function coupled with a bit of gravity information to mesure the direction pointed. So the user can simply point into a direction and start flying.
A couple of days before my final presentation at school I asked around for datasets to play with and present. I actually got 10! Thanx again guys! These are some screenshots of these datasets in the final application.
