Data science insights and the classification of terrorist attacks in Nigeria using machine learning techniques

Abstract

Terrorism has become a critical concern, with extremist groups orchestrating widespread violence resulting in casualties, displacement, and societal instability. Despite growing interest in counterterrorism, limited research has applied data science to classify and interpret patterns of terrorist activity. This study addresses this gap by employing data driven approaches to analyze terrorism in Nigeria. Utilizing data from the Global Terrorism Database and the Armed Conflict Location and Event Data Project, the study conducts exploratory data analysis to uncover patterns and trends in terrorist incidents. The analysis, performed using Power Business Intelligence, reveals key insights into the distribution, characteristics, and relationships within the datasets. Machine learning classifiers including Decision Tree, Random Forest, and Logistic Regression are trained and evaluated using metrics such as accuracy, precision, and recall. Among the models tested, Random Forest demonstrated the highest accuracy. The findings reveal a consistent pattern of successful attacks and underscore the potential of data science techniques in understanding and predicting terrorist activities. This research highlights the value of analytical methodologies in supporting informed decision making and developing strategic interventions for counterterrorism efforts.

Keywords: Armed conflict location; decision tree; event data project; machine learning classifiers; random forest.