Defining Interaction Techniques for Visualizations on Touch-based Devices.

Team Members: Ramik Sadana, John Stasko

    Expanding Selection for Information Visualization Systems on Tablet Devices - ISS 2016
    Designing Multiple Coordinated Visualizations for Tablets - EuroVis 2016
    Designing and Implementing an Interactive Scatterplot Visualization for a Tablet Computer - ACM AVI 2014
    Interacting with Data Visualizations on Tablets and Phones: Developing Effective Touch-based Gestures and Operations (poster) - IEEE InfoVis 2013
    TouchViz: a case study comparing two interfaces for data analytics on tablets - ACM CHI 2013

    VIDEO: Expanding Selection for Information Visualization Systems on Tablet Devices, ISS '16 (92 MB mp4)
    VIDEO: Multiple Coordinated Views on a Tablet, EuroVis '16 (60 MB mp4)
    VIDEO: Touch techniques for the scatterplot, AVI '14 (86 MB mp4)

This research project explores the convergence of two technologies: information visualization and multi-touch display devices. Tablet computers have had a widespread acceptance and there has been a tremendous growth in the use of machines where touch is the primary interaction paradigm. Unfortunately, current information visualization systems translate poorly to these touch displays because data representations are typically small, especially for large data sets, and they commonly employ many fine-grained WIMP-style interface control components such as scrollbars, sliders, check boxes, dialogs, and text fields. These small data elements and interface controls are difficult to accurately select, manipulate, and control via fingers and touch.

Our objective in this research has been to analyze, understand, and develop a better set of interactions for information visualization on touch-based devices. This research contains significant design and evaluation components as we develop new methods of interaction for these applications. We have created a system called Tangere that runs on Apple iPads. It provides multiple coordinated views including scatterplot, line graph, histogram, and parallel coordinates visualization techniques. Interactions with data in any one view are reflected in the others as well. The system allows control of view placement and provides dynamic query controls.

In more recent work, we added the capability of various advanced selection techniques to Tangere. Through the use of the non-dominant hand's thumb as a kind of "clutch", we allow the set of interactve touch gestures to be extended. The videos and publications above illustrate Tangere's operations and capabilities.

Ultimately, we aim to contribute to a "science of interaction" in the field of data visualization as we gain a stronger understanding of how interaction contributes toward knowledge acquisition and insight generation.

This research has been supported in part by a Google Faculty Grant and by the National Science Foundation via award IIS-1320537.