Chapter 13 - Geovisualization

Don't worry if you're not clear about distinctions between digital cartography, computer mapping, GIS, etc, because there are no fixed boundaries between these concepts. But until electronic devices duplicate the portability and dependability of sheets, a fundamental distinction will remain between atoms v. bits (paper v. screen). I like to turn questions around: What if humans had first to understand Mars by visiting there before it was seen and mapped from afar; and what if Earth had been discovered with a telescope and successively remotely sensed before it was visited, etc. Similarly, how would geography have evolved if we had begun with GIS before paper maps?

13.1 Introduction: uses, users, messages, and media

Academics may be thought of as peaceful people, yet both this chapter and the last begin with discussions of military maps. In fact, cartography has probably always been led by military needs, and this continues to be true today, even though we can't look very far into the current US military agency responsible for "mapping." Another leader has always been the British Ordnance Survey whose very name is militaristic.

We use computers to map because this is in many ways an advance beyond paper, as illustrated by Figure 13.3, which contrasts a 1776 paper map with a digital map of the same area. I'm not sure why the scales of the 3 maps (paper, outline, and yellow land/blue ocean) are different, but you can certainly see that the computer map offers such advantages as flexibility and ease of updating. Yet some will claim that 300 years of technical "progress" have not produced a more attractive visualization than the paper version.

Figure 13.4 MacEachren is indeed a great source of innovation in cartography, one of the first scientists to set forth a rigorous treatment of uncertainty. (And don't forget his look-alike GIS guru Scott Morehouse, whom you may remember from Chapter 7.)

Personally I regard the operations of query and transformation as being fundamental GIS activities and not just a part of 'geovisualization,' but so be it: these need to be somewhere in the book. But look at our syllabus and note that there is also a full GTKAGIS chapter devoted to query.

13.2 Geovisualization and spatial query

The geovisualization progression presented at the beginning of this section can be a useful way to guide our understanding of GIS. Perhaps an example (based on the steps introduced in this section) will help:
  1. Exploration: You map land cover for the Chesapeake Bay watershed to see whether land cover data georegisters with county boundaries and water quality sampling stations.
  2. Synthesis: You explore whether forest lands have been converted to agriculture in certain counties, perhaps near the Bay or its tributaries.
  3. Presentation: You make a simple map showing the above associations (land cover, water quality), with a legend, scale bar, and locator map.
  4. Analysis: You use spatial regression to predict whether water quality changes are indeed predicted by landscape changes.
Figure 13.5 is a 3-dimensional visualization of these operations, although their order is rearranged. I find that the figure makes more sense if it's turned upside down, with public/low-interaction/data sharing at the bottom and specialist/high-interaction/knowledge construction at the top. Again, think of the hierarchy:
knowledge > information > data

Make a note to come back to the questions in this section because they comprise most of the important queries we make in GIS. For those who are mathematically inclined, question #5 ("Where has...occurred") corresponds to determining the set of places x where Z are found
{x: f (x) = Z}
You did this kind of operation when you delineated land below sea level in GTKAGIS Exercise 6b. There is also a useful set of examples in the last paragraph of this section.

And think a bit about how GIS - or any other complicated computer application - might have evolved had we not become stuck with WIMP so early on! Certainly this interface keeps us more distant from the map than did the old paper and pen interface.

13.3 Geovisualization and transformation

In my Insights to Section 8.1 I presented a simple metamodel of how models build on the real world, and Figure 13.6 is the authors' similar view. Table 13.2 gives a few examples of transformations that are part of this process, but I prefer a more complicated approach. My own treatment of transformations illustrates several ideas:
  1. Data represents phenomena of different dimensions, usually integral but also
  2. Fractal (non-integral) that can be
  3. Transformed into one another by using various GIS tools.
In fact, almost all of GIS consists of these kinds of transformations, which usually require parameters and can be composed. If f( ) and g( ) are two transformations, they may be applied serially to provide a new transformation:
G(z) = g(f(z))
E.g. you might represent a region by its centroid and then fit a surface to the points, as in a technique known as "kriging" (look it up in the text). Or you might make Voronoi polygons out of the regions and then locate the vertices of the polygon edges. Even if you don't follow this, appreciate that it keeps an army of programmers busy.

One reason I find the term "geovisualization" somewhat inadequate is that these transformations frequently provide statistical information that may be at least as useful as the cartographic product.

Another transformation technique discussed in Section 13.3.3 is dasymetric mapping, used to distribute mass from one layer according to information from another. But beware that the use of the word "incidence" does not correspond to its epidemiological use. (Incidentally, one has to love these neologisms: one of my favorites is Waldo Tobler's "pycnophylactic interpolation"!)

13.4 Immersive interaction and PPGIS

I agree that public participation in GIS is critical, for several reasons. It gives communities insights into and even command of the tools often used by more powerful, wealthy, and  technologically sophisticated groups. It may also give community members opportunities for careers in a rapidly growing industry. And at a minimum it may improve data. But I suspect that the evidence for the transformative power of GIS is limited. (An early example of this kind of thing is the work of Kevin Lynch, whom I knew at MIT in the 1970s and who invited people to create mental maps of their neighborhoods. And even before that my fellow UC Berkeley planning students were using technology to foster community organization in Oakland CA.)

An example of very sophisticated compositional transformation (see above) is shown in Figure 13.17. where a flat image is draped on elevation data. Technically the result is less than 3-dimensional (and sometimes called 2½D) because the resulting objects do not actually occupy free-floating volumes.

13.5 Consolidation

This chapter helps take us from simple flat maps, through sophisticated transformations and analyses, to the creation of virtual reality. As a career perspective the opportunities are quite broad, from computer cartography, through GIS technologies and medical imagery, to computer graphics imagery in films. I suspect than some of you are in this class in order to take advantage of new opportunities.

Perhaps it's my age, but looking at Figure 13.19 I sometimes think that, as our virtual worlds become more exciting, our real world may become more boring. But then the virtual world is becoming the real world...