Data visualization helps human beings gain insights into data and communicate their findings. However, most current visualization creation tools either only support fixed sets of designs or require an in-depth understanding of programming concepts. To enable general audiences to create custom visualizations for data-driven storytelling, we design interactions and implement user interfaces for visualization authoring. In the first part of the dissertation, we introduce and evaluate a series of three visualization authoring tools using traditional user interfaces: (1) iVisDesigner, which uses a data-flow model and enables users to author visualizations by specifying mappings from data to graphics interactively. (2) ChartAccent, a tool for annotating a given visualization; and (3) Charticulator, which allows users to design custom layouts interactively. In the second part, we extend our approach to multiple presentation media or display environments, including traditional 2-dimensional screens, large projection-based virtual-reality (VR) systems, and head-mounted virtual/augmented reality displays (HMDs). To leverage such immersive visualization environments, we ported and extended the iVisDesigner authoring approach to projection-based virtual reality. To facilitate the development of immersive visualizations, we built a visualization library called Stardust, which provides a familiar API to utilize GPU processing power in a cross-platform way. Using Stardust and incorporating ideas from iVisDesigner and Charticulator, we present a novel system for authoring data-driven stories in virtual and augmented reality.
Date:Monday, December 4, 2017 - 8:00am
Title:Visualization Authoring for Data-driven Storytelling
Committee:Tobias Hollerer (Chair), Matthew Turk, George Legrady, Bongshin Lee