Department of Information Science
Heinrich Heine University
Country-specific Sentiment on Microblogs
Can expressed sentiments (positive, neutral, negative) within text messages of microblogging services be differentiated by the users’ country? Are there indeed country-specific emotional patterns of microblogging behavior? Using the example of the service Twitter, the work in hand intends to introduce a measure for the expressed sentiments of countries. This idea assumes that people of different regions and cultures have a specific emotionality, not only in daily life but also in writing microblogs. 14,192,182 tweets have been analyzed from different countries, languages, and topics to be able to get significant results in how distinct people represent their positive and negative feelings and opinions to the outside depending on their home and culture. We applied a lexicon-based approach and adjusted the term lists (words as well as emojis) intellectually. As indicators we introduced sentiment strength (general emotionality), sentiment climate (difference between positive and negative sentiment values) and the amount of positive as well as negative microblogs. Finally, we compared our findings with other indicators on country level (as Human Development Index or average temperature).