Clustering Tweets via Tweet Embeddings

Abstract

Twitter is a popular social media platform where users interact through follows and tweets. This work explores computational methods of analyzing tweets with regards to understanding users and their interests. We consider various embedding models to produce tweet embeddings, which we then use to cluster the tweets, forming groups of semantically similar tweets. We then compare these tweet clusters to users clustered by interest based on accounts they follow. This work introduces techniques on how to effectively cluster tweets by semantic meaning despite the colloquial structure of tweet language. We also discuss how the topics of these tweet clusters align with the interests derived from the follow-based clustering approach, and provide insights into where they do and don’t intersect.