The Dynamics of Ukraine-Russian Conflict through the Lens of Demographically Diverse Twitter Data
Salim Sazzed
Abstract
Due to the disastrous global impact of the recent Russia-Ukraine conflict, people all over the world are getting concerned- expressing their opinions and feelings toward it on social media such as Twitter and Facebook. To comprehend people’s perceptions and reactions, in this study, we analyze around 80000 demographically diverse tweets representing Twitter users of 7 countries across four continents. We first perform sentiment analysis to reveal how sentiments are expressed across demographically diverse tweets. We notice a similar pattern in the conveyed opinions and sentiments across the groups, constituting a high number of negative tweets citing the disastrous consequence of the war and criticism of the aggression. In addition, we investigate the presence of highly offensive words and phrases in various demographic groups. Finally, we employ two topic modeling approaches, LDA and transformer-based BERTopic, to find the most prevalent and discussed topics across users of demographic groups. We witness the presence of a diverse set of topics in the tweets of various groups, with dominant themes conveying support for Ukraine, the destructive impact of the war, and the resolution of this conflict. Our analysis reveals the feelings, thoughts, and sentiments of demographically diverse Twitter users regarding this conflict and provides insights for understanding people’s perceptions across the globe.