Projection-based Visualization of Dynamical Processes on Networks

Abstract

Dynamical processes on networks play an important role in systems biology and statistical physics. To understand these processes, it is essential to understand which topological properties of the network are the main factors for the dynamics. We present a visualization approach that allows for such investigations. For visual encoding of the network topology, we use a node-link diagram. The nodes, however, are placed according to the dynamical processes. We use a projection method of the time series data to generate animations that maintain the mental map and exhibit the behavior of the dynamics. Suitable coloring schemes for the nodes encode the current values of the dynamics and individual nodes can be investigated with linked views to a time series plot. We present case studies to demonstrate that our approach is effective for the observation whether the dynamical processes follow the network topology.