Tweet sentiment as proxy for political campaign momentum

Social media has become very popular over the past decade. There are millions of users across the world sharing information with each other instantaneously through several social media platforms. With these many users sharing huge volumes of data analysis of social media data has become a prominent area of research. Recent studies on the use of data from social media platforms such as Twitter for predicting political elections have raised many questions as well as created the interest in using Twitter data for predictive analysis. The overarching objective of this paper is to study the capability of Twitter data as an ex-ante indicator of event outcomes by modeling the momentum of political campaigns. Three indicators with predictive capability are proposed as measures of momentum of political campaigns. An asset price model is adapted to model momentum of candidates. Empirical validation is provided based on Twitter data from the 2014 US midterm election and the 2016 Presidential primary elections. Our results support the argument that data from social media can be considered as a reliable predictor of events in political campaigns.