Sentiment Analysis of Tweets to Gain Insights into the 2016 US Election

Abstract

Social media use is at an all-time historic high for the United States, so we considered one popular social media platform, Twitter, and tried to see if we could predict how a group of people felt about an issue by only using posts from social media. For our research, we looked at tweets that focused on the 2016 United States presidential election. Using these tweets, we tried to find a correlation between tweet sentiment and the election results. We wrote a program to collect tweets that mentioned one of the two candidates, then sorted the tweets by state and developed a sentiment algorithm to see which candidate the tweet favored, or if it was neutral. After collecting the data from Twitter and comparing it to the results of the Electoral College, we found that Twitter sentiments corresponded with 66.7% of the actual outcome of the Electoral College. The overall sentiment of all tweets collected leaned more positively towards Donald Trump than it did for Hillary Clinton. Using the data that was collected, we also looked at how different geographical locations affected a candidate’s popularity, analyzed what issues were most prevalent in tweets, and looked at the ratio of a state’s population versus the number of tweets gathered.