Exploring Food Safety Emergency Incidents on Sina Weibo: Using Text Mining and Sentiment Evolution

Abstract

Food safety remains a crucial concern in both public health and societal stability. In the age of information technology, social media has emerged as a pivotal channel for shaping public opinion and disseminating information, exerting a substantial influence on how the public perceives incidents related to food safety. This study specifically focuses on the “Rat-Headed Duck Neck” incident as a case study, conducting a comprehensive analysis of extensive social media data to investigate how online public discourse molds perceptions of such events. To accomplish this research, data were initially gathered using a custom web crawler technology. These data encompassed various aspects, including user interactions, emotional expressions, and the evolution of topics. Subsequently, the study employed an innovative approach by combining BERT-TextCNN and BERTopic models for a thorough analysis of sentiment and thematic aspects of the textual data. This analysis provided insights into the intricate emotions and primary concerns of the public regarding incidents related to food safety. Furthermore, the research harnessed Gephi, a network analysis tool, to scrutinize the dissemination of information within the network and to monitor dynamic shifts in public opinion. The findings from this study not only shed light on the role of online public sentiment in the propagation of food safety events but also provide fresh perspectives for policymakers and business leaders when responding to similar crises, taking into account the subtleties of online public sentiment. These innovative methodologies and findings significantly enhance our comprehension of public responses to food safety incidents within the realm of social media.