This mark is automatically generated based on current and recent activity in the IV/LAB.

* The number of lab members who have logged in to our linux machines in the past 24 hours is 0.
* The number of lab members who have logged in to our linux machines this week is 0.
* The current size of the directory for our published research papers is 24150055 KB.
* The lights in the lab are on.

Visualization-by-Sketching: An Artist's Interface for Creating Multivariate Time-Varying Data Visualizations

Journal Article

Visualization-by-Sketching: An Artist's Interface for Creating Multivariate Time-Varying Data Visualizations. David Schroeder, Daniel F. Keefe. IEEE Transactions on Visualization and Computer Graphics (2015) Volume 22, Number 1 pp. 877-885

Notes

Best Paper Award, Also presented at the Best of IEEE VIS session at SIGGRAPH 2016

Abstract

We present Visualization-by-Sketching, a direct-manipulation user interface for designing new data visualizations. The goals are twofold: First, make the process of creating real, animated, data-driven visualizations of complex information more accessible to artists, graphic designers, and other visual experts with traditional, non-technical training. Second, support and enhance the role of human creativity in visualization design, enabling visual experimentation and workflows similar to what is possible with traditional artistic media. The approach is to conceive of visualization design as a combination of processes that are already closely linked with visual creativity: sketching, digital painting, image editing, and reacting to exemplars. Rather than studying and tweaking low-level algorithms and their parameters, designers create new visualizations by painting directly on top of a digital data canvas, sketching data glyphs, and arranging and blending together multiple layers of animated 2D graphics. This requires new algorithms and techniques to interpret painterly user input relative to data “under” the canvas, balance artistic freedom with the need to produce accurate data visualizations, and interactively explore large (e.g., terabyte-sized) multivariate datasets. Results demonstrate a variety of multivariate data visualization techniques can be rapidly recreated using the interface. More importantly, results and feedback from artists support the potential for interfaces in this style to attract new, creative users to the challenging task of designing more effective data visualizations and to help these users stay “in the creative zone” as they work.

Video
Projects

This publication is a part of the following research projects:

Links