Mapping Associations - Mapping Scientific Frontiers: The Quest for Knowledge Visualization - page 101

Tabl e 3. 3 Comparisons of
layout- and texture-based
visualizations

Number of images 279

Links in PF by layout 319

Links in PF by texture 284

Common links 1

Expected common links 2:34

Point probability 0:23

Information 0:14

3.2.4 Visualizing Abstract Structures

Information visualization has a long history of using terrain models and relief maps
to represent abstract structures. Information visualization based on word frequencies
and distribution patterns has been a unique research branch, especially originated
from information retrieval applications.

3.2.4.1 ThemeView

The changing patterns at the lexical level have been used to detect topical themes.
Some intriguing visualization technologies have been developed over the past few
years (Hetzler et al. 1998 ).
The most widely known example in this category is ThemeView, developed at
Pacific Northwest National Laboratory (Wise et al. 1995 ). James Wise described
an ecological approach to text visualization and how they used the relief map as a
model of a thematic space (Wise 1999 ). ThemeView enables the user to establish
connections easily between the construction and the final visualization. Figure 3.9
is a screenshot of PNNL's ThemeView, showing word frequency distributions as
peaks and valleys in a virtual landscape.

3.2.4.2 VxInsight

Sandia National Laboratory developed a visualization system called VxInsight to
model clustered information in the form of a virtual landscape. It adapts the popular
landscape model to visualize underlying data. In particular, researchers at Sandia
National Laboratory used VxInsight to visualize cluster structures derived from
Science Citation Index (SCI). VxInsight allows the user to move back and forth in
the virtual landscape. Figure 3.10 shows a virtual landscape produced by an earlier
version of VxInsight.
VxInsight was applied to the analysis of patents (Boyack et al. 2000 ). Thematic
terms and patenting companies are cross-references in landscapes over a few periods
of time by labeling key thematic terms and coloring different companies. Figure 3.11