Sentiment Analysis of Text Using Twitter Data

• In Event: Putting Text Together: Criminal Justice Applications of Text Analytics

Abstract

The millions of tweets shared on Twitter daily provide a rich resource of public sentiment on thousands of topics. In the wake of highly publicized officer-involved shootings, many people took to social media to express their opinions, both positive and negative, of police. This presentation will explain how to conduct text and sentiment analysis using millions of tweets that mention the words “police” or “cops.” We will walk through how to systematically process millions of tweets, and several methods of sentiment analysis. The text analysis methods will include cleaning tweets, stemming words, creating document-term matrices, and extracting info within tweets, such as the number of emojis. Sentiment analysis methods will range from simple bag-of-words classification to more complicated machine learning prediction. We will also discuss the challenges and benefits of sentiment analysis and working with large amounts of text data available through Twitter.

Authors

• Ashlin Oglesby-Neal, Urban Institute

• Emily Tiry, Urban Institute

• KiDeuk Kim, Urban Institute