Fall 2015
Project of CS 579: Online Social Network Analysis, taught at the Illinois Institute of Technology by Aron Culotta.
Project Introduction: It is believed that public sentiment is correlated with the behavior of the stock market. In a famous paper, Bollen et al. (2010) made the claim that Twitter mood is correlated with the Dow Jones Industrial Average (DJIA), and that it can be used to forecast the direction of DJIA changes with 87% accuracy. Besides its obvious significance in investing, this surprising result challenges several fundamental notions in the social sciences, such as the efficient market hypothesis.
In this project, I verify whether the surprising results of Bollen can be reproduced and whether they can produce a profitable investment strategy. Unfortunately, I find that measuring Twitter mood does not offer an improvement over a learning algorithm that only uses past DJIA values.