With a plethora of candidates running in 2016. We were wondering how predictable the current state of media politics can be.
What it does
So by combining leading candidate data from Twitter and a clever sprinkling of algorithm magic we have been able to simulate how candidates today use social media to direct their strategy.
How We built it
Using Markov Chains and a twitter scraping tool we were able to extract and analyze a large data-set of tweets which are then fed to a live-stream front-end that simulates the current political climate.
Challenges I ran into
Half of the group was inexperienced with working with web tools. In addition we had a different implementation in mind but the API we were using was too young into production and thus was not capable of servicing our business object.
Accomplishments that we are proud of
Despite the late pivot we were able to quickly distill what is the core function behind the app.
What we learned
Markov Chains are cool - like lifesaving awesome content cool. IBM Bluemix was awesome (we hope to use it in the near near future)
What's next for House of Bots
Create your own celebrity bot! Twitter flame war simulator and possibly a tinder for cross-fit uber drivers(.xyz)