Ukrainian Crisis on Twitter: Sentiment Analysis - Possible Interpretations

Abstract

The mass use of social media has enabled users to reach out and share opinions,attitudes, and emotions on various topics - current events in particular - almost instantly.In this paper, we analyze the attitudes of Serbian speakers toward Russia and Russiansduring the ongoing war in Ukraine, based on the material obtained using an applicationfor collecting and processing comments on Twitter. The study covers the period fromFebruary to September 2022. Starting with the assumption that Twitter users tend tobe freer and more spontaneous in expressing their views compared to users of othersocial media (Facebook, Instagram, etc.), about 11,000 tweets were collected using theBigBoxData application based on keywords (Russia, Ukraine, war, special operation,Putin, Zelensky...), and were subsequently manually filtered and annotated. Thegoal was to find out if and how the attitudes of Serbian speakers toward Russia andRussians have changed with the ongoing war in Ukraine compared to associations andstereotypes toward Russia and Russians reported in previous papers. The combinationof qualitative (positive, negative, and neutral sentiment) and quantitative (percentageshare of each of the three sentiments) sentiment analysis showed that the attitudes ofSerbian speakers toward Russia and Russians have changed compared to the previousperiod (using as a reference the Associative Dictionary of the Serbian language and theReverse associative dictionary of the Serbian language, both of which report extremelypositive associations). The methods used in this study and the research results canserve for further research and attitude change within the crisis discourse on socialmedia and the Internet in general