Visual Analytics For Studying Complex Networks


Wednesdays@NICO Seminar, Noon, May 19 2010, Chambers Hall, Lower Level

Prof. Takashi Nishikawa, Northwestern University


The network-oriented view of complex systems has become wide spread over the last decade or so, mainly driven by statistical analyses that detect specific connectivity structures/characteristics, such as small-world and scale-free properties, as well as communities of densely connected nodes. Given a real network data set, however, one often do not know a priori what type of structures/characteristics to look for, and therefore need to try to search and discover something unknown. Going beyond the problem of detecting a given structure, I will discuss our approach for tackling more challenging problem of discovering unknown structures in complex networks, based on the concept of visual analytics, an integrated framework for analyzing high-dimensional data sets with a combination of scientific visuali-zation and user interactions. As a first example of this approach, I will describe a visual and interactive method for discovering distinct groups of nodes in a network using a user- selected set of node properties computed from the connectivity structure. Our method has potential for discovering the com-munity structures, k-partite structures, or more general group structures in which the groups can be dis-tinguished by a combination of node properties. I will demonstrate that our method can effectively find and characterize a variety of group structures in model and real-world networks.