ACLED Clustering & Cell Networks

By Robert Franssen in R clustering geostatistics

August 8, 2021

Applying spatial clustering and geostatistical techniques with interactive mapping to explore the relationships between armed-conflicts.


This analysis seeks to identify space-time clusters of paramilitary violence against civilians and explore their relationship with cellular communications infrastructure. It serves a broader conversation regarding what makes communities susceptible to the rapid recurrence of violence and whether access to communications infrastructure enables communities to avoid subsequent attacks from paramilitary actors by serving as an early warning network. This analysis is focused on armed-conflicts occurring in the Eastern Democratic Republic of the Congo along the edge of Lake Kivu, including Virunga National Park. However, the input data and code allow this analysis to be re-run for any extent geographic extent in Africa.

This inspiration for this analysis extends from a series of attacks perpetuated by the Lord’s Resistance Army (LRA) between 24 December 2008 and 13 January 2009. Over this two-week period, LRA members attacked several neighboring communities killing 620 civilians and abducting more than 160 children. The violence was exacerbated by communities' inability to coordinate or communicate warnings to neighboring areas. As a result, it took several days for the United Nations Mission in Congo (MONUC) to learn of the violence [@http://zotero.org/users/local/yMlW3Shk/items/4UA8ICCV]. Communities elsewhere without access to telecommunications may be at a similar increased risk of experiencing progressions of violence.