Analyzing Sentiment and Themes in Fitness Influencers' Twitter Dialogue

摘要

Social media allows anyone to distribute content and build an audience. Natural language processing, sentiment analysis, and psycholinguistic text analysis have proven to be powerful tools for characterizing and classifying social media text. Furthermore, the combination of text and sentiment analysis have allowed researchers to identify influencers both by their structural roles and the content they produce. In this paper, we investigate fitness-oriented social media influencers. This research aims to understand how fitness influencers (N = 92) on Twitter speak to their audiences through thematic and sentiment analysis of their tweets (N = 273,868). Findings suggest sentiment and topics discussed vary between male and female health and fitness influencers on the platform. The analysis also determined no sentiment differences between self-identified fitness trainers/coaches and influencers who do not identify as such. The results have implications for personalization and recommendation algorithms that operate in this space.