Meme it Up: Patterns of Emoji Usage on Twitter

Abstract

As emojis have grown in their popularity in social media over the last decades, they not only enrich messaging with emotional connotations but offer a convenient system for studying the effects of memes on ideas’ survival. In this work, we treat emojis like standardized memes to test the impact of their usage on different facets of success within social media. Specifically, we extracted random individual tweets from Twitter to construct a list of emojis used within each tweet. With this dataset, we aimed to address three distinct questions: (1) whether there are specific patterns of emoji usage that increase tweet popularity; (2) whether emojis usage on tweeter can be a good predictor of the stock market trading volume; and (3) whether there is a specific subset of emojis associated with low-quality tweets (e.g., spam). We found no evidence of the positive effects of emoji usage on tweet popularity. However, there was a reason to claim that negative emojis may trigger an intensive response from the audience. For some companies, we were able to accurately predict stock patterns based on emoji usage. Finally, there clearly was a specific subset of emojis used in low-quality tweets. This work may serve as a starting point for a deep investigation of the emoji-meme system, as this topic seems to be relatively fresh in the literature.