Networked data mining based on social network visualizations

Studies in social network theory focus on characterizing complex social relationships by firstly mapping and visualizing them into a graph, and then subsequently identifying the corresponding graph properties. This paper provides an integrated approach, which combines social network analysis and data mining theory with the necessary geographical attributes to analyze 1417 instances of terrorism that occurred world wide during the period 1980-2002. The study reveals interesting patterns on the evolution of these terrorist organizations over two decades. The proposed method can be easily generalized to be applied to other types of large-scale networked datasets, such as micro-array data, and genomic networked data, etc.