Facebook. com is the place where people share their preferences in pop culture. In this project we tried to use some of information of the user to see how the data taken from the profile can describe a person. And the main personality trait that we wanted to examine is how mainstream is the facebook user according to the information we pull from his/her profile. We chose movie page and tv show page likes in our application. And depending how popular are those movies and tv shows on interned give each person an index indicating his level of “mainstreamness”. So just getting that index would not satisfy a viewer so we decided to give each user who logs in a fish, and put it into the pig pool with other users’ fish. And depending how mainstream is the user the behavior of fish would change. In our case less mainstream fish would move faster and its range of movement is higher indicating how far that fish is from the mainstream culture.
When a user logs into facebook via (umichparty.com/art/login.html)(running on firefox is recommended) we get their ‘likes’ for tv shows and movies. With this information we use AJAX to query an imdb api server to get that movie or tv show’s rating and number of votes. We then append this data to a file to a remote server.
The processing application then downloads this file from the remote server and parses it into a data structure that the application can use. Each movie or tv show is given an initial percentage with votes*rating/(votes of the most popular movie or tvshow*rating of the most popular movie or tvshow). This gives us an absolute value for how mainstream someone is but we want it to be relative to everyone in the pool so we calculate a new relative mainstream value using (absMainstreamValue-minAbsMainstreamValue)/(maxAbsMainstreamValue-minAbsMainstreamValue). This way, the most mainstream person gets a value of 100% and the most hipster gets 0% and everyone else who is in between those values.
The relative mainstream value for each user corresponds to the speed and radius of the movement of the fish in the stream. If the user clicks a location on the screen, the fish will move toward that spot.
We are very glad with the result we got. We think it is quite interesting to see a representation of your interests in the pool with some other people and see how mainstream you are relatively to other users. The graphics might look a little simple, but we think it looks nice and overall pleasing to the eye. Of course our project can be developed further and the range of the data used for determining the fish behavior could be wider.
Here is the video link:
3 fish example:
many fish example: