A Text Mining Analysis of Academic Libraries' Tweets

This study applies a text mining approach to a significant dataset of tweets by academic libraries. The dataset for this research was collected from the complete Twitter timelines of ten academic libraries. The total dataset comprised 23,707 tweets with 17,848 mentions, 7625 hashtags, and 5974 retweets. Academic libraries from the dataset have typically posted fewer than 50 tweets per month, though tweet volume grew rapidly in late-2013 through 2014. The results show variance between academic libraries in distribution of tweets over time. The most frequent word was “open,” which was used in a variety of contexts by the academic libraries. It was noted that the most frequent bi-gram (two-word sequence) in the aggregated tweets was “special collections”. The most frequent tri-gram (three-word sequence) was “save the date”. The most frequent word categories in the semantic analysis for most libraries were related to “knowledge, insight, and information concerning personal and cultural relations”. The most common category of the tweets was “Resources” among all the selected academic libraries. These findings highlight the importance of using data- and text-mining approaches in understanding the aggregate social data of academic libraries to aid in decision-making and strategic planning for patron outreach and marketing of services.