Better Intelligence Through Microblogging: Text Mining Tweets for Actionable Information
Croll, Meghan · Carolina Digital Repository (University of North Carolina at Chapel Hill) · 2019
The goal of this paper is to evaluate specific event-related Twitter posts, called tweets, for their usability as intelligence data. Using text mining software for Natural Language Processing to attempt to establish the best keyword search strategy for actionable information, 2000 tweets documenting 2 independent events were analyzed. Additionally, the Twitter users who provided the tweets were evaluated for source reliability to determine if their information was likely to be credible. The results of this study suggest that the best information retrieval strategy for tweets featuring reliable and actionable information, is to use hashtagged topics and/or location details as search keywords. Not only is "location" frequently a facet of determining actionability, the co-location of event and user increases the likelihood of a high-reliability source.