The novel coronavirus pandemic (COVID-19) that started in 2020 became a global public health emergency. It hit the world economy heavily and it continues to impact people’s lifestyles in every country around the globe. Due to the pandemic preventive and control measures such as avoiding gatherings, reducing travel and even the closures of city passages and country borders, social media became an important channel for people to exchange information on pandemic.
Analysis of public opinions on pandemic is an important foundation for emergency response strategy formulating and adjustments. It plays a significant role in making precise and differentiated strategies for all or key areas.
This study was based on the source data from the social media Sina Weibo. It analysed the public opinions on COVID-19 during the early stages of the pandemic outbreak based on space, time and content. First, it conducted time series analysis and spatial feature distribution on COVID-19 related posts. Topic extraction and classification model was constructed based on the LDA topic model and Random Forest algorithm using stratification method to refine gradually from broad towards specific. 9 first level topics and 17 second level topics were identified from the text content of COVID-19 related posts.
