Inspired by Biology: Towards Visualizing Complex Networks

Abstract

The term “biological network” comprises a large and multifaceted set of different types of networks. These different network types bring with it unique visualization and visual analysis challenges. We first survey the literature in order to characterize and identify outstanding gaps in the visualization of biological networks. Inspired by these many challenges and difficulties faced by the field. Specifically, we focus on three challenges of particular interest to us: i) improving the visual quality of commonly employed straight-line node-link diagrams, ii) the visualization of uncertainty in networks, and iii) the visualization of group structures in compound graphs. To tackle these three challenges, we conduct five investigations.To tackle challenge 1, we first investigate the principled and algorithmic splitting of vertices to iteratively resolve edge crossings and thereby improve the readability of graphs. As an alternative solution to challenge 1, we investigate the visualization of so-called ego-networks, which allow for the visualization of only node-relative and node-relevant topology, instead of the entirety of a network. Third, within the context of challenge 2, we investigate the visualization of node attribute uncertainty using animated “wiggliness”, i.e., animated node motion. Fourth, in order to tackle challenge 3, we survey the current state of compound graph visualization and, finally, we combine the aforementioned four works together and develop a prototypical dashboard for the visualization of compound graphs.