This original research article applies computational methods from network science and event data analysis to model the structural dynamics of ceasefires and conflict recurrence in South Sudan. Using a novel dataset synthesised from the Armed Conflict Location & Event Data Project (ACLED) and the South Sudan Conflict Database, the study constructs temporal networks of armed groups and analyses the impact of ceasefire agreements on localised violence patterns. The results quantify the fragility of peace arrangements, identifying key network configurations—specifically, factional fragmentation and alliance instability—that significantly predict renewed hostilities. The discussion situates these computational findings within the political economy of South Sudan's peace processes, arguing for data-driven early warning systems. The conclusion outlines implications for conflict mediation and the integration of computational social science into African peace and conflict studies.