happy not sad, sad not happy
I’ve set up two twitter bots that follow identical algorithms but will probably evolve to be very different. Both @happyB0T and @sadB0T are learning Markov text generators. Their vocabulary starts from zero but they gradually expand it by searching for tweets and adding them to their own growing dictionary of words and phrases. They then use a Markov chain algorithm to construct new sentences and tweet them.
The difference between the two is that happyB0T only searches for tweets that include :) (happy faces), and sadB0T sadly only collects tweets that include :( (sad faces).
More specifically in code and twitter api search terms…
searchQuery = “:) -:(“; //happy, not sad
searchQuery = “:( -:)”; //sad, not happy
The bots also serve as fragmented mirrors to our cultural and linguistic tendencies. Only time can tell how these two will progress.
A project done in Processing with Twitter4J and RITA libraries, by Matthew Plummer-Fernandez.
Note: the ‘0’ in happyB0T and sadB0T is actually a zero 0. (For twitter naming purposes)