A Little Birdy Told Me: Analysis of the Impact of Public Tweet Sentiment on Stock Prices

The combination of the advent of the internet in 1983 with the Securities and Exchange Commission’s ruling allowing firms the use of social media for public disclosures merged to create a wealth of user data that traders could quickly capitalize on to improve their own predictive stock return models. This thesis analyzes some of the impact that this new data may have on stock return models by comparing a model that uses the Index Price and Yesterday’s Stock Return to one that includes those two factors as well as average tweet Polarity and Subjectivity. This analysis is done with ten different large, public firms on the NASDAQ and NYSE. Our results suggest that models that implement the Twitter data perform slightly better than their classical counterparts