Spatio-temporal evolution of public opinion on urban flooding: Case study of the 7.20 Henan extreme flood event

Abstract

Under global climate change, the increasing frequency of urban floods caused by heavy rainfall has attracted widespread public attention. Simultaneously, social media serves as an emerging data source, playing an important role in monitoring and researching disaster-related public opinion. In this study, we took the 2021 7.20 Henan extreme flood event as an example, extracting highly relevant microblogs from China's ‘Weibo’ platform to explore the spatio-temporal evolution of public opinion. Firstly, a quantitative period division index was constructed based on the relationship between rainfall and Weibo activity, divided the study period into four parts. Using the LDA ((Latent Dirichlet Allocation) model, we extracted topics and used sentiment dictionary analysis to extract emotions. Subsequently, we analyzed Weibo activity, topics, and sentiment across different periods and conducted comparative assessments. There was a response relationship between precipitation and the number of microblog posts, which varied across different periods. Early on, anxiety and panic regarding the flood surfaced during disaster warnings. During the disaster phase, people in affected areas primarily expressed panic, aid expectations, and needs. Those in surrounding regions exhibited concern for the flood, empathy towards affected people, and support for relief efforts. While after relief operations had commenced, public emotions experienced a rapid recovery. This study presents a quantitative method for dividing flood disaster periods and a comprehensive analysis approach considering both temporal and spatial dimensions to explore public opinion dynamics in typical flood events. It offers suggestions to relevant authorities for implementing targeted emotional management measures and mitigating disaster impacts.