Feature-Based Visual Analytics for Studying Simulations of Dynamic Bi-Stable Spatial Systems

Abstract

Simulations of dynamic bi-stable spatial systems usually generate large and complex data that are hard to evaluate. In this paper, we describe how visual analytics technology can help in analyzing such simulation data. The idea behind our approach is to utilize concepts of feature-based visualization. Consequently, we consider (1) interactive specification of meaningful features, (2) analytic extraction and tracking of features as well as detection of events in the features' evolution, and (3) visual representation of features with their spatial, temporal, and structural aspects. Our solution has been used by simulation experts to analyze spatio-temporal distributions of multiple types of particles in reaction-diffusion simulation data. With the help of the feature-based approach the scientists were able to understand how the spatial separation of proteins develops over time.