Evolution of online public opinion during meteorological disasters

Meteorological disasters are important public events that can generate a wide range of online public opinion. Studying the patterns and mechanisms of public opinion dissemination during meteorological disasters and moderately strengthening the voice of official media can alleviate public nervousness and facilitate disaster prevention, reduction, and recovery. Therefore, taking Typhoon Mangkhut as an example, we collected data from Sina Weibo in China and Twitter in the Philippines. Based on a 'data preparation-public opinion mining-data analysis' framework, patterns and characteristics of the evolution of public opinion were identified through social network analysis and sentiment analysis methods. The results showed that public opinion surrounding Mangkhut differed in the two countries. The trend in public opinion was 'low-high-low.' During natural disasters, shifts in opinion exhibited a 'positive-negative-positive' pattern. In the Philippines, netizen sentiment reached lowest point 24-48h after the typhoon landed and recovered steadily and quickly. However, among Chinese netizens, sentiment hit lowest point later, mostly because of a man-made negative event. To help people cope with natural disasters, the Chinese official media should promptly release accurate information, play a more active role in guiding public opinion, and pay more attention to man-made negative events during disasters.