"The central aim of this work is to explain what causes the various types of political violence found within and across African states. In contrast to the widespread view that conflict is confined to a few crisis prone states, new evidence suggests that almost all states are sites of substantial, widespread political insecurity (Raleigh et al. 2010). Civil war accounts for 35% of the conflict across African states; the remaining 65% is composed of communal and political militia violence, rioting, protests and violence against non-combatants outside of a war context. These forms of ‘invisible’ violence often involve state collusion and present a widespread risk to civilians. This proposal takes a holistic approach, looking at all forms of political violence and seeking to explain them within a novel theoretical framework emphasizing two stages of onset indicators, and employing the latest available disaggregated data methodologies for spatial and temporal dynamics. Further, it introduces spatial and scaled approaches, which are the most rigorous and well suited to a comprehensive conflict study, as risks, triggers and dynamics are spatially inscribed and hierarchical. The theoretical contribution of this work is an examination of how insurgency and opposition violence are spatial and political processes that are shaped by the political, economic and social geographies of states. The empirical contributions include an extension of the most comprehensive data on political events (ACLED) and a merging of these disaggregated data with information on local level political, economic, social and environmental conditions throughout Africa. Methodologically, this project is the first to test spatial and temporal forecasting methods on real-time conflict hotspots. Finally, the conclusions of this work have applicability to development agencies, governing bodies and international regimes concerned with the growing threat emanating from failed and failing states."