MotionGlyphs : Visual Abstraction of Spatio-Temporal Networks in Collective Animal Behavior

Abstract

Domain experts for collective animal behavior analyze relationships between single animal movers and groups of animalsover time and space to detect emergent group properties. A common way to interpret this type of data is to visualize it as aspatio-temporal network. Collective behavior data sets are often large, and may hence result in dense and highly connectednode-link diagrams, resulting in issues of node-overlap and edge clutter. In this design study, in an iterative design process, wedeveloped glyphs as a design for seamlessly encoding relationships and movement characteristics of a single mover or clustersof movers. Based on these glyph designs, we developed a visual exploration prototype, MotionGlyphs, that supports domainexperts in interactively filtering, clustering, and animating spatio-temporal networks for collective animal behavior analysis. Bymeans of an expert evaluation, we show how MotionGlyphs supports important tasks and analysis goals of our domain experts,and we give evidence of the usefulness for analyzing spatio-temporal networks of collective animal behavior.