Study on Spatio-Temporal Topic-Sentiment Synergy Model and Visualization of Online Public Opinion on Public Health Emergency

Abstract

Public health emergencies can generate online public opinion on social media platforms such as Weibo. Existing studies show that both temporal and spatial factors have an impact on public opinion, topic and sentiment mining of public health emergency microblogs can realize the monitoring, prediction and guidance of public opinion considering the temporal and spatial factors. Taking the outbreak period of the Delta variant in three different regions of China in 2021 as the research object, this paper constructed a model based on the Latent Dirichlet Allocation (LDA) model, improved SnowNLP lib and sentiment map. Data processing, topic mining and sentiment calculation were carried out to realize the synergistic analysis of topic and sentiment. Results illustrate that this model can reveal the law of online public opinion evolution and sentimental intensity, that online public opinion is influenced by spatial and temporal factors, especially that small cities with smaller volume and attention need to be focused on the observation and guidance of public opinion.