Public Attention about COVID-19 on Social Media: An Investigation based on Data Mining and Text Analysis

The COVID-19 epidemic is influencing global population. Social media has become important platforms to acquire and exchange information during the outbreak of COVID-19. This study explores public attention on social media. Popular Weibo texts related to COVID-19 with coronavirus and pneumonia as the keywords during December 27, 2019 and May 31, 2020 were collected in our study for public attention analysis. By combining data mining and text analysis, the public attention level trend in different stages were presented. Then a correlation analysis between public attention level and COVID-19 related cases number, topic analysis, and sentiment analysis were conducted. Significant positive correlation between public attention level and COVID-19 related cases number was identified. Based on Latent Dirichlet Allocation model, topic extraction was implemented in different stages and 41 topics were identified totally. For a comprehensive understanding of public emotions, sentiment analysis was performed. This study provides valuable lessons for public response to COVID-19.