Visualizing uncertainty in node-link diagrams - A user study

Abstract

Uncertainty visualization is manifold and applied in many disciplines of visualization. In node-link diagrams, the edges can have uncertain attributes; these are to be presented directly in the graph. On the one hand, the uncertainty values are encoded with visual variable. On the other hand, the user has to decode the visualization to identify the original value. In this paper, we focus on four uncertainty visualization techniques and their suitability to enable the user to decode and identify the correct values. In our evaluation, we investigate the maximum number of different levels that can be used for each technique such that a user is able to reliably distinguish different values.