A Tweet-Ranking System Using Sentiment Scores and Popularity Measures

Abstract

Various fields look to Twitter as a marvelous repository of users’ opinions. Decision-makers rely on Twitter to examine tweets about various topics of interest. In doing so, they can determine the nature, context, and accompanying emotions of what people are discussing. The topics taken up in such discussions include current affairs, events, and products, and tweets can indicate both the intensity of discussions and how people react to them. In this paper, we propose a system that collects tweets and applies sentiment analysis to them. It allows users and decision-makers to browse an ordered and categorized list of tweets based on their sentiment and popularity measures, such as retweet and favorite. An essential advantage of our system is that it can be used more generally to explore any topic of interest with which the public is engaging on Twitter. Our findings show that using the number of retweets and the number of favorites enhances the sentiment scores and enables the highlighting of tweets’ most common opinion on a given topic.