Daisy Timms Undergraduate Dissertation 2015/16

The TWITTERATI: The Predictive Potential of Sentiment Analysis and Twitter

Supervised by M.Hepple

Abstract

Over 500 million tweets are broadcast daily on Twitter. 

The volume of information is inconceivable to a mere human, however by using a tool such as sentiment analysis, defined as the extraction of opinions or emotions from raw data, we can attempt to derive meaning from it.

This project will explore the potential of sentiment analysis on Twitter data in order to predict the public response to a public event, notably to forecast the outcome of a reality television series. This provides an apt domain for experimentation in this field, as there are multiple public votes throughout a short window of time.