Geography Reference
In-Depth Information
(Fanea, Carpendale and Isenberg, 2005), the added complexity of the adapted versions
means that the upgrade from 2D to 3D may not provide as much benefit as expected. The
pseudo-3D versions of bar and pie graphs which appear increasingly in research publications
introduce apparently unnoticed distortions into the messages conveyed to readers.
10.2.2 Displaying additional data variables
An immediate gift of the third display axis is that it enables at least one additional data
variable to be mapped during visualization. At the simplest level, this means that the height
axis may be used for showing the values of any interval-scaled variable. This technique has
been used (arguably to mixed effect) in the prism map, which overcomes the inability to
apply the size visual variable to the surface area of polygons without distorting the area of
real-world features. (2D cartograms solve this problem in another way, but at the expense
of introducing spatial distortions to well-known maps that are often confusing to data
analysts.)
However, it should be noted that the third display axis is not restricted to displaying a
single additional data variable, as in the prism map. One way of maximizing the potential of
the
z
-axis is to map several data variables onto more complex 3D symbols or glyphs. Several
examples have been discussed in the literature, and they generally fall into two groups. The
first group involves the construction of complex glyphs from multiple data variables, which
are typically distributed across the
x
-
y
plane of the display space. These glyphs are most
commonly used in information and scientific visualization, a representative form being a
tree whose branches are sized, angled and coloured to represent particular data variables.
Analysts interprets the distribution of such glyphs in much the same way that they would
explore the distribution of any set of point symbols.
Where data are available for points on an equally spaced grid, an alternative approach has
been adopted. In the iconographic display (Levkowitz, 1991; Pickett and Grinstein, 1988),
data values control the display of an array of articulated icons across the data grid, whose vary-
ing orientations, thicknesses and colours result in spatial textures that vary visually across the
grid. The interpretation of these textures is undertaken differently from the more discrete dis-
play of 3D glyphs, because the iconographic display is designed to be perceived preattentively
rather than cognitively. Although this display was first devised using 2D icons, later versions
have been developed for both 3D and 4D visualizations. For example, the 3D approach has
been used to display multiple environmental variables in North America by Healey (1998).
As we will show later, however, even imaginative use of multivariate 3D glyphs and icons is
insufficient for the visualization of datasets containing large numbers of variables.
10.2.3 Providing a familiar view of the world
Two-dimensional representations of three-dimensional features in the real world have sig-
nificant interpretive drawbacks. By flattening topography and using abstract symbols to
represent surface features, the 2D paper map imposes a significant learning burden on oc-
casional map users. (In a similar way, the traditional architectural drawing, with its attempt
at capturing 3D reality in a multiple view - plan, side elevation and front elevation - divides
visual attention and makes spatial integration more difficult.) By contrast, it is often claimed
that viewers find it easier to interpret data visualizations that are constructed as 'natural'
Search WWH ::
Custom Search