Quasi-cliques are a type of dense subgraphs that generalize the notion of cliques, important for applications such as community (or module) detection in social (or biological) networks. While quasicliques are originally defined only for undirected graphs, we have recently generalized the concept to directed graphs by proposing (𝛾1,𝛾2)-quasi-cliques, which have density requirements in both inbound and outbound directions of each vertex in a quasi-clique subgraph. This generalization broadens application since many real graphs are directed such as communication or citation networks, and gene regulatory networks. We also proposed efficient parallel algorithms with advanced pruning techniques to find maximal (𝛾1,𝛾2)-quasi-cliques, paving the way for interactive mining (e.g., by tuning parameters such as 𝛾1 and 𝛾2) and result examination. In this demonstration, we present a novel graphical user interface on top of our mining program, called QCQ-Viewer, for users to (1) mine maximal (𝛾1,𝛾2)-quasi-cliques of an input graph with an interactive interface easy for parameter tuning, and to (2) examine the resulting dense subgraphs in an aesthetically-pleasing manner with the help of graph visualization functionalities

Click on the image below to open the YouTube video