Exploring Twitter Messages during the COVID-19 Pandemic in Sri Lanka: Topic Modelling and Emotion Analysis

The tremendous growth of social media empowers the rapid dissemination and amplification of information by becoming the most influenceable communication tool with a significant extent of human lives. People extensively use those platforms during the outbreak of the COVID-19 to cope with social distancing and isolation by expressing their opinions, views, and issues they are facing. The study used Twitter as the social media platform to discover the emotional perspective of Sri Lankans during the outbreak. The emotion analysis was conducted based on 8-scale emotions (anger, anticipation, disgust, fear, joy, sadness, surprise, and trust) provided by the NRC Emotion Lexicon. The extraction of noteworthy topics propagated in different domains of COVID-19 helps to identify the perception of society during the pandemic. Latent Dirichlet Allocation was applied to explore relevant and accurate topics. Furthermore, the wordcloud was generated by representing the most insightful data points which help to keep track of social concerns or viewpoints. Our experiment shows that people discuss under three major themes of "COVID-19 Immunization", "Government rules and regulations to prevent the spread of COVID-19" and "COVID-19 related health statistics". Furthermore, the "fear" is dominated among eight emotions followed by "anticipation" and "sadness" when Sri Lankans express their thoughts and feelings related to COVID-19. There is a noticeable decrease in "joy" which is the least expressed emotion followed by "trust". Hence, clearly defined, these findings support in predicting the COVID-19 related outcomes and addressing the concerns, emotions, and reactions of citizens by the government and the policymakers.