Data-Driven Analysis of Public Concerns in Major Construction Accidents and Intervention Strategies

Major construction accidents not only cause property damage and casualties but also trigger public panic and media crises, necessitating precise emergency management strategies. Existing research often relies on qualitative analysis and does not fully consider the differences between accident types and the public’s concerns. Based on the ConCA model, this study integrates machine learning and big data analysis to explore the impact of different accident types on public attention and proposes differentiated emergency response strategies for each accident type. Public concerns were categorized as rational, emotional, or moral by analyzing social media comments. The study found significant differences in public concerns across accident types: human factor accidents primarily evoke both rational and emotional concerns, natural disasters predominantly elicit emotional concerns, and equipment defects mainly lead to rational concerns. In response, this study suggests three differentiated intervention strategies: a “high responsibility + high attention” strategy for human factor accidents, a “high responsibility + moderate attention” strategy for equipment defects, and a “low responsibility + low attention” strategy for natural disasters. This study fills the gap in public cognitive behavior models by proposing an emergency management framework based on accident types and public concern patterns, which provides important guidance for improving emergency response efficiency and public trust in the future.