Analyzing conversations with dynamic graph visualization

In this paper, we consider the problem of analysis and visualization of online conversations (chat histories, email archives, etc.). We present a dynamic graph drawing algorithm based on modification of multidimensional scaling. The algorithm builds a layout of sequence of graphs and produces a slice view of the evolution of online communications. The method have been applied for visualization of two real-world datasets. We show how to use these visualizations for analyzing and extracting hidden temporal patterns from online conversation data.