CAREER: Information Visualization Research and Education

Suny At Buffalo, Amherst NY

Investigators

Abstract

The research goal of this project is to develop more effective techniques for visualizing information that will facilitate and enhance information-processing tasks. In particular, the project develops four novel techniques, for: (1) constructing better graph visualizations, (2) detecting clusters in a database using visualization, (3) constructing visualizations faster using previsualizations, and (4) compressing 3D models. The graph visualization technique uses a novel graph-decomposition-based approach to construct more effective graph visualizations, which can help the user to more easily understand the information represented in a graph. The cluster detection technique allows the user to visually detect clusters in a database by visualizing its contents, and can be a useful aid in datamining applications. The previsualization-based technique can significantly reduce the average time needed to construct visualizations in interactive applications, where visualizations are repeatedly constructed, by storing precomputed and previously computed visualizations in a previsualization cache, and checking the cache first before constructing a new visualization from the scratch. This technique can help in developing visualizations systems with better response times. The technique for compressing 3D models uses the concept of canonical ordering to efficiently compress 3D polygonal models, and reduce their space requirements and transmission times over a network. The educational goals of this project consist of teaching a course on information visualization, developing the corresponding course-material, advising graduate students, mentoring undergraduate students and initiating them to research in information visualization. The results of this project will benefit researchers and users in the areas of datamining, computer graphics and interactive visualization, and in applications such as software engineering, where information can be represented using graphs.