Anon Tweets

For this project I wanted to parse the chaotic nature of Anonymous , by compiling a list of every Anonymous twitter account I could find (~400, although they are always branching off) and plotting their influence vs. signal noise on a graph. The points furthest away from the origin (0,0) and in the middle are most influential with the most information output. To add a more immediate informational point, I mapped the size and color of each object based on the number of re-tweets of the latest tweet.

The X axis = # of Tweets for that account

The Y axis = # of Followers for that account

Size/color of Spheres = Amount of re-tweets their latest tweet received.

If one clicks on a member, they are shown their last tweet and the twitter account name.

I ran into some problems early on, with the nature of the Twitter API. I was only allowed 350 query requests total per hour, so not even one instance of the program could parse the accounts, so I figured out to build a list a members and query that. Unfortunately, Twitter only allows you to pull 20 users from each list, so I had to make 20 lists of 20 members.

I then ran into a problem with the way Twitter wrote “retweet_count” into JSON – if it is over 100 tweets, it outputs “100+” which screws up the JSON library parsing. I had to go into the library and hack it apart…

I also need to refine the scalability of the program, as the range of followers/tweets spans from < 100 to > 30,000 – for now, I am mapping smaller values proportionally in a higher range than larger values, so the smaller points aren’t condensed in the corner.

For the next step, I want to refine the visual aspects of the program, somehow show the gradual shift in influence, by requesting new data from the server in a set amount of time (not over the twitter limit) and then synching it with the NY times API feed to show the level of Anonymous twitter activity in relevance to news cycles.



Related projects:

Ant Colony
Mobility Shifts: SMS Game
Ad Hoc Architects
Social Cooking App