This is a project we started as part of a Hackathon during Spring 2016 at Berkeley, CA.
Our goal is to try to understand how the public feels about presidential candidates. We will collect tweets about candidates and try to compare them based on the proportion of positive and negative tweets and perform some permutation tests to see who is the most likeable candidate. We will try to compare Bernie Sanders, Hillary Clinton and Donald Trump.
Results:
Based on data collected on May 17th: Sanders beats Clinton and Trump. Trump beats Clinton.
All comparisons presented high levels of significance.