Visual Analytics and Imaging Laboratory (VAI Lab)
Computer Science Department, Stony Brook University, NY
Abstract: We propose the concept of teaching (and learning) unfamiliar visualizations by analogy, that is, demonstrating an unfamiliar visualization method by linking it to another more familiar one, where the in-betweens are designed to bridge the gap of these two visualizations and explain the difference in a gradual manner. As opposed to a textual description, our morphing explains an unfamiliar visualization through purely visual means. We demonstrate our idea by ways of four visualization pair examples – data table vs. parallel coordinates, scatterplot matrix vs. hyperbox, linear chart vs. spiral chart, and pie chart vs. tree map. The analogy is commutative – any member of the pair can be the unfamiliar visualization. A study we conducted suggests that this new paradigm can be an effective teaching tool. We found that for all of the four pairs we studied users could understand the unfamiliar visualization method either fully or at least significantly better after they observed or interacted with a series of transitions from the familiar counterpart. Our examples provide good insight how effective visualization pairings can be identified, and we hope that they will inspire other visualizations transformation pairs and associated transition strategies to be identified.
Teaser: Illustrating the visualization method of parallel cordinates by morphing from a table.
The presentation of data via a table is a very familiar paradigm to most people. On the other hand, data presentations via parallel coordinates is far less commonly known. Here we see our system in action. It starts off with a data table and then transforms it into a parallel coordinates plot, one dimension at a time. Two highlighted samples are also shown to provide a frame of reference and enable a training by example experience. Users can either watch the sequence as an animation or manipulate the transitions via a slider.
Watch this video to see how it works:
Video: Next is a video that shows the other pairs we tested -- scatterplot matrix vs. hyperbox, linear chart vs. spiral chart, pie chart vs. tree map, and data table vs. parallel coordinates.
Demo: Try it yourself -- click here
Paper: P. Ruchikachorn, K. Mueller, "Learning Visualizations by Analogy: Promoting Visual Literacy through Visualization Morphing," IEEE Trans. on Visualization and Computer Graphics, 29(9):1028-1044 2015. ppt pdf