Direct Manipulation Techniques for Creation of Multiple-View Visualizations

Abstract

I propose an extension to Lyra, a tool for authoring interactive data visualizations, that introduces novel processes for creation and laying out of multiple-view (MV) visualizations. The existing tools for designing data visualizations lack specific support for MVs, making the process of creating such visuals cumbersome for users. I draw on the concepts of direct manipulation, in which users interact with objects in the visual to enact desired changes, to develop new methods of creating MVs that are more intuitive to users and have less of a learning curve. With my proposed changes, users can iteratively add new groups in Lyra simply by dragging and dropping visual marks onto dropzones. To facilitate this new method for adding groups, I implement a layout concept within Lyra to allow for easier creation of gridded layouts for MVs. Additionally, this concept is extended to data-driven layouts that can be automatically generated for small multiples, or trellis charts. The system is evaluated by demonstrating the new mechanisms and showing an example gallery of MVs that can be created with these methods.