A Data Science Workflow for Discovering Spatial Patterns Among Terrorist Attacks and Infrastructure
Daniel C. Fortin, Thomas Johansen, Samrat Chatterjee, George A. Muller, Christine F. Noonan · 2021
Terrorism continues to plague nations around the globe. Decision makers and analysts need data-driven tools to help them gain insight into terrorist groups, uncover trends, and quantify the risk of terrorist attack. We introduce an interactive data visualization application to explore incidents of terror at user-specified spatial and temporal levels using data from the Global Terrorism Database. The application allows a user to view historical terrorist activity on an interactive world map, with features that allow filtering and visualization in multiple forms. Additionally, we discuss a statistical modeling approach to determine the relationship between terrorist attacks and types of infrastructure within a country using zero-inflated models for count data. The modeling framework is demonstrated using case studies on terrorist attack data from Libya and the Jammu and Kashmir region in India. The models are able to identify statistically significant infrastructure variables and identify specific regions of interest for further investigation by analysts.