Geography Reference
In-Depth Information
data, which 3D techniques are most appropriate for particular applications, and when might
2D approaches be more appropriate? (Indeed, is 3D always better than 2D?) Third, what
can we learn from other communities in which 3D graphics and visualization technologies
have been developed? And finally, what are the key R&D challenges in making effective use
of the third dimension for visualizing data across the spatial and related sciences? Answers
to these questions will be based on several lines of evidence: the extensive literature on data
and information visualization; visual perception research; computer games technology; and
the author's experiments with a prototype 3D data visualization system.
2
10.2 What is gained by going from 2D to 3D?
There are several good reasons for using the third dimension when visualizing data. Some
of these are briefly reviewed below.
10.2.1 Additional display space
However large the monitor, and however high the screen resolution, users will always reach
limits to the amount of data that can be reasonably displayed in their data visualizations.
Research in the field of information visualization (or infovis) indicates that the volume of
data objects that can be comfortably displayed on screen is considerably larger when using
a 3D representation than when using a 2D representation (Card, Robertson and Mackinlay,
1991). Where information exists in tabular form, then conventional 2D box displays may
be converted to 3D solids to increase the information shown on screen; and where the
information involved is hierarchically organized, such as the folders and files stored on a
computer, or the documents available in an online repository, then conventional 2D trees
may be upgraded to higher capacity 3D cone trees (Robertson, Mackinlay and Card, 1991).
Finally, where data are organized as a network, then 2D connectivity graphs may be replaced
by 3D node-and-link graphs (Hendley
et al.
, 1995). A major advantage of this increase in
information for the user is that it permits a larger amount of contextual information to
be seen while focusing on particular objects of interest. A significant drawback is that 3D
visualizations often tend to impose severe interaction demands on users.
The conversion of existing 2D data visualization techniques into 3D equivalents is an
attractive development strategy, and many conversions have proved successful, as is the
case with 3D fish-eye distortion displays (Carpendale, Cowperthwaite and Fracchia, 1997),
3D beamtrees (van Ham and van Wijk, 2003) and 3D distribution glyphs (Chlan and
Rheingans, 2005). However, other kinds of 2D visualization techniques may not deliver
equally significant benefits when converted into the third dimension. For example, although
software has been developed to generate 3D versions of parallel coordinates and star glyphs
2
The experimental 3D data visualization software used to generate the illustrations that accompany this
chapter was devised by the author, based on data visualization principles and the published results of exper-
imental research in visual perception and human-computer interaction. The software was programmed by
his son Iestyn Bleasdale-Shepherd, who is a software engineer specializing in real-time computer graphics
at Valve Corporation in Seattle.
Search WWH ::
Custom Search