Abstracts:Abstract Public construction projects often encounter social conflicts when project complexity increases and diverse stakeholder concerns arise, such as environmental, economic, and route-selection issues. Social conflicts that originate from diverse interests disrupt the planned project execution, resulting in cost overruns and schedule delays. However, research on up-to-date measures to investigate and monitor such social conflicts remain insufficient. To address these challenges, this study proposes a network-based social conflict analysis model for public construction projects using news articles. The proposed model comprises two main functions: conflict identification and assessment. To identify conflicts, this study employed a large language model to select relevant news articles and extract project stakeholders as well as conflict-related issues. The identified stakeholders and issues were used to construct a two-mode network. Subsequently, the proposed model was used to perform a network analysis to identify key stakeholders and conflict issues as well as to calculate the complexity of conflicts. In addition, conflict severity was determined according to rules established by the authors based on the decision tree method. Next, the model was used to evaluate the overall conflict level based on two dimensions: conflict complexity and severity. Lastly, the proposed model was applied to two construction projects in South Korea that have experienced significant social conflict. The findings confirmed that the model sufficiently reflects the progression of actual conflicts. Therefore it is expected to enable up-to-date investigation of social conflicts and facilitate efficient conflict management during project execution by tracking the dynamics of conflicts and providing quantified conflict levels.