A rapid disaster intensity assessment method using social media data: a case study of the flood disaster in the Beijing-Tianjin-Hebei region

With the worsening of climate change, extreme weather events are on the rise, leading to more frequent occurrences of climate-related disasters. Analyzing people's perceptions and attitudes towards disasters after they occur can help determine the spatial pattern of the disaster intensity and the post-disaster needs of different populations. The implication is to provide references for disaster assessment and post-disaster relief needs analysis.

 

Starting from July 29, 2023, due to the influence of Typhoon Dusrayi and Typhoon Canu, the Beijing-Tianjin-Hebei region in China suffered from catastrophic rainfall, resulting in severe flooding in multiple areas. This study utilized web crawlers to collect relevant Weibo data during the disaster, applied machine learning models to conduct public opinion analysis on the flooding disaster, developed the evolutionary patterns of public opinions on the disaster, and obtained heat maps and sentiment indicators for different cities. The results will contribute to the rapid assessment of post-disaster losses and guide the resource allocation in the initial emergency rescue process after the disaster.