Online community users’ perceptions of particulate matter in South Korea through topic modeling and semantic network analysis

In this study, we investigate public perception of particulate matter (PM) in South Korea to address the persistent high public anxiety levels surrounding this environmental issue. For this purpose, we first compared the monthly trends of PM concentrations, news articles, and online community posts. Furthermore, we performed a cross-correlation analysis between PM concentrations and online community posts, which are considered to be the most public-aware source, to examine how the public's response changes with variations in PM concentrations. Additionally, we conducted topic modeling and identified nine topics based on 2714 online community posts from January 2013 to December 2023. The results showed that public concern about PM is related to PM concentrations but media was found to play a role in amplifying it. The most prominent topic was the public’s critical attribution of PM in Korea to China. Other topics included interest in PM forecasts, physical risks and personal responses to PM, requests for more active government response, and controversy over the South Korean government’s PM reduction policies. In addition, the public has significant health concerns, leading to individual protective actions such as mask and air purifier purchases, possibly influenced by media and marketing strategies. There is a significant gap between public perceptions of the causes of PM and the direction and content of government policies. For instance, although the public tends to attribute PM primarily to external factors such as pollution from China, government policies appear to emphasize internal sources, such as regulating vehicle emissions. The discrepancy was found to make the public feel dissatisfied with policy implementation. This study demonstrates how the public perceives and interprets PM issues. It is also significant that we provide empirical evidence of the relationship between PM concentrations and online community posts through cross-correlation analysis. In this study we advocate for the need for policymakers to communicate PM-related information effectively to facilitate public alignment with PM reduction policies.