Analysis of GIS Data by Simple Examination - Introducing Geographic Information Systems with ArcGIS - page 375

Exercise 6-2 (Project)

Categorization and Symbolization
Thinking about Maps Again

Now that the connection between the table and the map is firmly re-established in your mind, we
turn back to the map itself. You could broadly classify maps into two categories: general-reference
maps and thematic maps. Many maps, like road maps or topographical maps are used for general
information, navigation, topology, and to show locations of features. Thematic maps (sometimes called
statistical maps), on the other hand, show polygons that divide up the landscape into distinct areas;
the areas are coded—with color or other symbolization—to indicate the value of a characteristic or
attribute about each individual area. So, in a sense, attribute information is reflected directly, visually,
on the map; in GIS you find this information mostly in tables. (Description of areas somewhat true for
general-reference maps as well, but in thematic maps this information is the primary reason for the
existence of the map.)

Classification (or Categorization) and Symbolization
Basic statistics is one way of understanding bunches of numbers. Another way is to place things
in categories, based on some pertinent number associated with each thing. For example, we could
categorize (i.e., place into separate groups) the students in a class according to their heights. Put
everybody less than 4 feet tall in group A. Those 4 feet or more but less than 4 feet, 3 inches belong
in category B. Those who are 4 feet, 3 inches or more but less than 4 feet, 6 inches go in category C,
and so on. So here's the general concept: We have k objects and we place each of them in one, and one
only, of n categories, where n is less than (or, in a trivial case, equal to) k. Almost without saying, the first
category consists of a set of smallest numbers, the next category consists of the set of next smallest of
numbers, and so on.

Consider another example: Suppose that we had a set of numbers (which I have put in order to make
things simpler).

1 1 2 3 3 4 5 7 8 10 10 11 12 15 19 19 22 23 25

That is, we could place all the numbers from low to high, say, in a text string from left to right.

If we wanted three categories we might partition them as follows by assigning obvious breaks between
categories:

1 1 2 3 3 4 5 7 8 10 10 11 12 15 19 19 22 23 25

Category 1 Category 2 Category 3

Here the simplicity ends. The goal is to arrange things so that humans can best understand the nature
of whatever is being studied. To do this, three fundamental questions must be addressed: (1) How many
categories are appropriate for a given set of data? (2) What approach should be used to partition the
objects into categories? (3) How might the results be effectively presented or displayed.