Junior Management Science
Social media platforms as enabler for real time and many-to-many communication play an important role in the analysis of consumers’ opinions, attitudes, moods, and behaviors towards brands. Emojis as a non-verbal, explanatory and emotional component are increasingly used for a more expressive online communication. While current emotion mining tools only focus on text analysis, we are the first who conduct an automated sentiment analysis of brand-related tweets containing emojis in addition to text. Within the scope of our master thesis at the marketing faculty at University Duisburg-Essen we analyzed 999,197 Starbucks-related and 566,597 McDonald’s-related tweets. We used tweets directed at two different global brands in the fast food sector to increase generalizability. On a sentiment polarity scale, the analyzed tweets show a rather positive
sentiment value towards Starbucks and a slightly negative sentiment value towards McDonald’s. We also find that sentiment is classified identically across brands for 94% of emojis. We conclude that the sentiment value can be considered as an indicator for the perceived image of a brand. Our approach provides an innovative tool for companies to directly analyze emotional content on social media platforms and improves the understanding for the needs of consumers.
Keywords: Sentiment analysis, Emoji, Twitter, brand, Stimmungsanalyse