Date of Submission
Campus Only Senior Thesis
Bachelor of Arts
Professor Angela Vossmeyer, Ph.D.
2020 Spencer T Sheff
A new trend called “nowcasting” is growing popular amongst investors. Nowcasting is a term used to explain models trying to accomplish very short-range forecasting using real time data. An enormous source of real time data is the social media platform Twitter, where people constantly “tweet” about anything ranging from random thoughts during the day to their political beliefs and more. There has been increasing literature in the field of economics on relationships that social media platforms share with the stock market’s performance. The use of social media to gauge investments has gained so much popularity that JP Morgan created an index called the “Volfefe” index which tracks President Donald Trump’s tweets.
This study builds onto the literature by highlighting statistically significant relationships found between tweets and the stock market. To show these relationships, we first use natural language processing algorithms and data mining techniques to gather and analyze the sentiment of over 3,500,000 tweets pertaining to twelve companies in the S&P 500. We then derive a metric called “public opinion” from tweet sentiment that serves as a proxy for latent utility in the market that investors expect to gain by owning equity in a firm. We use public opinion to model stock price, returns, and volatility. The result is a statistically significant relationship between public opinion and our three response variables. We then test the predictive power of public opinion, and find evidence supporting that a metric of this type can be useful as a tool in a predictive model.
Sheff, Spencer, "Telling the Whole Story: How Natural Language Processing and Crowdsourced Information Can Give Us a New Perspective into Financial Markets and Latent Utility" (2020). CMC Senior Theses. 2477.
This thesis is restricted to the Claremont Colleges current faculty, students, and staff.