The shape of Shakespeare: visualizing text using implicit surfaces
Randall M. Rohrer, David S. Ebert, John L. Sibert · 2002
Information visualization focuses on the use of visual means for exploring non-visual information. While free-form text is a rich, common source of information, visualization of text is a challenging problem since text is inherently non-spatial. The paper explores the use of implicit surface models for visualizing text. The authors describe several techniques for text visualization that aid in understanding document content and document relationships. A simple method is defined for mapping document content to shape. By comparing the shapes of multiple documents, global content similarities and differences may be noted. In addition, they describe a visual clustering method in which documents are arranged in 3D based upon similarity scoring. Documents deemed closely related blend together as a single connected shape. Hence, a document corpus becomes a collection of shapes that reflect inter-document relationships. These techniques provide methods to visualize individual documents as well as corpus meta-data. They then combine the two techniques to produce transparent clusters enclosing individual document shapes. This provides a way to visualize both local and global contextual information. Finally, they elaborate on several potential applications of these methods.