Exploring Eye Movements with Node-Link Graph Layouts

Analyzing eye movementdata is challenging, in particular, if visualizations are used that directly show the eye movementson the stimulus. However, it is important to identify visual scanning strategies in context of the semantics of the displayed stimulus. In this paper, we investigate the idea of providing node-link diagrams using multiple graph layouts in which each is directly linked to the visual stimulus, but represented in a separate view. Typically, a single graph layout provides specific informationand has certain benefits. Hence, we propose to compare several common and easy to implement graph layout algorithms in multiple-coordinated views to extract different scanning strategies of participants. The nodes of each node-link diagram hereby model areas of interest while the weighted links denote the transitions between the areas of interest. Additionally , we illustrate the usefulness of our approach by applying it to a formerly conducted eye trackingstudy focusing on route finding tasks in public transportation maps.