Gemini: A Grammar and Recommender System for AnimatedTransitions in Statistical Graphics
This addresses the challenge for visualization developers by reducing the effort required to create animated transitions, though it is incremental as it builds on existing perceptual studies and declarative approaches.
The authors tackled the problem of specifying effective animated transitions between statistical graphics by developing Gemini, a declarative grammar and recommendation system that automates design suggestions, achieving exact replication for 9 out of 11 designs and avoiding errors in participant implementations.
Animated transitions help viewers follow changes between related visualizations. Specifying effective animations demands significant effort: authors must select the elements and properties to animate, provide transition parameters, and coordinate the timing of stages. To facilitate this process, we present Gemini, a declarative grammar and recommendation system for animated transitions between single-view statistical graphics. Gemini specifications define transition "steps" in terms of high-level visual components (marks, axes, legends) and composition rules to synchronize and concatenate steps. With this grammar, Gemini can recommend animation designs to augment and accelerate designers' work. Gemini enumerates staged animation designs for given start and end states, and ranks those designs using a cost function informed by prior perceptual studies. To evaluate Gemini, we conduct both a formative study on Mechanical Turk to assess and tune our ranking function, and a summative study in which 8 experienced visualization developers implement animations in D3 that we then compare to Gemini's suggestions. We find that most designs (9/11) are exactly replicable in Gemini, with many (8/11) achievable via edits to suggestions, and that Gemini suggestions avoid multiple participant errors.