This study sought a programmatic way to analyze large bodies of Twitter data to find thematic and semantic trends. Many current approaches to Twitter data analytics rely on fairly basic searches, such as identification of particular hashtags. While these hashtags can be useful and descriptive when it comes to finding the threads of conversations revolving around a particular topic, not every Twitter user uses them, nor is the content of the tweet necessarily reflected by the hashtag used. The current approach sought to rectify this shortcoming, looking specifically at the body of the tweet and its language. As an example of this approach, we present our analysis of a body of tweets that center on the events in Ferguson, MO and the threads of discussion that we found. Additionally, we anticipate that what we’ve learned can be applied to any Twitter analysis, agnostic of topic or event.
Tanner Stirrat, ’15
Thao Nguyen, ’15
Ho Chi Minh City, Vietnam
Computer Science and Statistics
Sponsors: Ross Sowell & Ann Cannon