Bayesian spatiotemporal modelling of political violence and conflict events using discrete-time Hawkes processes

TL;DR

The paper addresses the need for rigorous, uncertainty-aware monitoring of political violence risk beyond simple historical averages. It develops a Bayesian spatiotemporal discrete-time Hawkes framework applied to ACLED data in South Asia, introducing two variants (nospatialSE and ST_SE) to capture regional baselines and spatial-temporal self-excitation, with inference implemented in Stan and a 52-week maximum excitation horizon. The authors demonstrate that the ST_SE model generally provides better fit across scenarios, yielding interpretable parameters for baseline risk and self-excitation, and offering practical tools for early warning, spatial risk visualization, and short-term predictions with quantified uncertainty. Compared to existing methods, the approach provides more stable risk thresholds and richer spatial-temporal insights, with potential for extension to covariates, multivariate interactions, and global application.

Abstract

The monitoring of conflict risk in the humanitarian sector is largely based on simple historic averages. The overarching goal of this work is to assess the potential for using a more statistically rigorous approach to monitor the risk of political violence and conflict events in practice, and thereby improve our understanding of their temporal and spatial patterns, to inform preventative measures. In particular, a Bayesian, spatiotemporal variant of the Hawkes process is fitted to data gathered by the Armed Conflict Location and Event Data (ACLED) project to obtain sub-national estimates of conflict risk in South Asia over time and space. Our model can effectively estimate the risk level of these events within a statistically sound framework, with a more precise understanding of uncertainty than was previously possible. The model also provides insights into differences in behaviours between countries and conflict types. We also show how our model can be used to monitor short and long term trends, and that it is more stable and robust to outliers compared to current practices that rely on historical averages.

Figures (13)

• Figure 1: Weekly events in each country for each conflict type.
• Figure 2: Proportion of total events of each conflict type by country.
• Figure 3: Total number of events over 5 year observation window. Black crosses indicate capital cities (including Karachi in Pakistan as the biggest city and economic capital).
• Figure 4: Proportion of weeks with zero events. Black crosses indicate capital cities (including Karachi in Pakistan as the biggest city and economic capital).
• Figure 5: Protests in Pakistan (unaggregated). Observed data on week $t$ (black bars) versus median of the estimated 50% quantile (red line) and median 95% interval of event counts calculated based on 100 posterior samples (red ribbon).