A Computational Framework for Analysing Conflict Narratives in South Sudan: A Natural Language Processing Approach to Peace and Conflict Studies

This original research article presents a novel computational framework for analysing conflict narratives in South Sudan, applying natural language processing (NLP) techniques to a corpus of local media reports and peace agreement texts. The study develops and validates a bespoke taxonomy for categorising conflict drivers and peacebuilding themes specific to the South Sudanese context. Quantitative and qualitative analysis reveals significant temporal shifts in narrative salience, particularly around land, ethnicity, and governance, correlating with key political events. The findings demonstrate the utility of computational methods in providing scalable, evidence-based insights for conflict analysis, offering a complementary tool for traditional qualitative peace and conflict studies. The framework's limitations and potential for real-time monitoring and policy analysis are critically examined.