Real-time emotional dynamics and public discourse on Sina Weibo following the first imported mpox case in mainland China: A sentiment analysis study.

Abstract

The human mpox has become a significant public health threat; however, little is known about real-time public emotional responses to mpox in China. We collected comments from the top 40 popular posts containing the keyword "Chongqing mpox" on Sina Weibo within 48 hours following the announcement of the first imported mpox case in mainland China (September 16, 2022). Using natural language processing (NLP), we applied the BosonNLP sentiment dictionary to assign emotional values (EVs) to each comment. Latent Dirichlet Allocation model was used to identify key public concerns, and Spearman correlation analysis was conducted to examine relationships between engagement metrics (comments, likes, responses) and EVs. The 40 popular posts received 5,107 comments, 144,940 likes, and 7,358 responses, respectively. Weibo users' comments reached a peak within two hours after the announcement (hourly data) and decreased markedly after four hours. Users with the most comments came from Guangdong, Shanghai, Beijing, Zhejiang, and Chongqing. The public was most concerned about attitudes and the transmission mode of the mpox. Sentiment analysis showed that the comments were mostly negative (76.13%). The EVs of the negative comments were 32.31 times higher than those of the positive comments. The numbers of comments, likes, and responses were all significantly associated with EVs (p < 0.001). Public sentiment following the first imported mpox case in mainland China was predominantly negative and highly concentrated within the first four hours. Findings underscore the need for timely, region-sensitive risk communication during emerging outbreaks.