Geography Reference
In-Depth Information
nCubes, there are five possible kinds of choropleth map: any cell as a percentage of the
universe; any cell as a percentage of the row total for one dimension; any cell as a percentage
of the row total for the other dimension; any row total on one dimension as a percentage
of the universe; any row total on the other dimension as a percentage of the universe. With
a three-dimensional nCube, there are 19 kinds of choropleth map, and we leave it as an
exercise for the reader to work out the permutations. Providing a user interface to select
between them is clearly non-trivial. A quite separate kind of option we hope to offer is
cartograms as base maps instead of conventional boundary maps, and even animating them
(Southall and White, 1997).
There are clearly large potentials for graphing higher-order nCubes, but these can only
be realized within a software environment that offers three-dimensional graphics and an-
imation. One aspect is that with one-dimensional nCubes there is usually one clear right
way of presenting the data; with higher dimensions there are often a series of possibilities
each of which will best reveal some facet of the data, and the user needs some quick mecha-
nism for moving between them. Animation is arguably essential, as the only way of clearly
presenting the time dimension, and as the web site's full name indicates, this resource is
all about visualizing statistics
through Time
. The potential for presenting changing age and
gender structure via a single animated population pyramid is very obvious. Some of the
infrastructure for such developments is already in place, as until the impact of the lottery's
technical rules became clear we planned to run GeoTools within users' browsers, and the
software hooks for doing this remain in the system.
However, there are other potentials which could only be realized using much higher
performance hardware. For example, the system holds geo-referenced historic maps for
different dates, and could be easily extended to include digital elevation model data for the
whole country. This would permit an interface in which users flew through a landscape
in which the scanned maps were draped over a relief model. Animated three-dimensional
graphics could then be located within this landscape to represent nCube structures associated
with each geographical area. With the exception of the DEM data, all the information and
metadata to support this vision is already within the system.
13.5 Conclusions
Academic visualization research is almost by definition about high-end visualization, be-
cause the low end is left mainly to the developers of Excel and various statistical packages.
However, the effect of this is to prioritise helping people with PhDs to better understand
their data over helping the general public, who perhaps need more help. Excel, in particular,
knows far too little about the data it holds to ensure that users only create graphs that make
some sense; the same is arguably true of the statistical maps that desktop GIS packages
can create. The enormous strength of the DDI-based visualization outlined here is that the
statistical metadata are not simply a source of labels for graphs, they capture the
logical
structure of the data, and constrain users' graphical choices to maps and graphs that make
some sense.
Academic researchers looking at the current main outcome of this research, the
Vision of
Britain through Time
web site, will be under-whelmed by the very unremarkable maps and
graphs it creates, and it has already been noted that we were heavily constrained both by
Search WWH ::
Custom Search