Topic Analysis of Social Media Posts during the COVID-19 Pandemic: Evidence from Tweets in Turkish

The aim of this study is to analyze the shift in the social media discourse during the COVID-19 pandemic. The sample included Turkish users on Twitter, who shared opinions about the pandemic between March 9 and October 31, 2020. The collected tweets were first classified with the Long Short-Term Memory (LSTM) architecture, which used the global vector for word representation embedding method. In addition, due to the grammatical and semantic structure of the Turkish language, we employed the Zemberek library for the text pre-processing stage. We analyzed data according to two categories: user-to-public posts and user-to-user posts. User-to-user data were investigated with effective social network analysis techniques. Empirical results showed that Twitter users posted and disseminated information mainly related to economy, politics and world topics.