Textbook: Slocum, T.
A. 1999. Thematic Cartography
and Visualization. Prentice Hall: New Jesey. See Page: 60-82
Data Classification
Data Classification: The purpose of classification is to simplify the
appearance of the map (maps with few classes 4-5 are easily understood).
Define classed and unclassed map?
Classed map is a map that displays classified data
(classes-portray relationship among the data). Unclassed map maintains correct
data relations (no classes). The use of classed or unclassed map depends on the
purpose of the map (e.g. to show spatial relationship, maintain correct data,
present, or explore data).
List and define the main methods of data classifications?
1. Equal interval: In this method each class occupies an equal
interval along the number line. The resulting equal map interval is easy for
map users to interpret and the legend contains no missing values or gaps. The
major disadvantage of equal interval is that the class limits fail to consider
how data are distributed along the number line (Figure 4.1 and 4.2 , page 64
and 65).
2. Quantiles:
In the quantiles method of classification, data are rank-ordered and equal
number of observations is placed in each class. Its advantages include class
limits can be computed manually, each class have almost the same area, it can
be used also for ordinal data (no numeric). Its major disadvantages are, gaps
result in a legend (difficult for reader to interpret) and it fails to consider
how data are distributed along the number line (Figure
4.1 and 4.2 , page 64 and 65).
3. Mean standard deviation: This method consider how data are distributed along
the number line. In this method classes are formed by repeatedly adding or
subtracting the standard deviation from the mean of the data (Table 4.4, page
69).
4. Maximum breaks:
This method looks at individual values along the number line and group those
that are similar to one another (avoid grouping values that are dissimilar).
However, the method miss natural clustering of data along the number line.
5. Natural break:
In natural break method graphs (e.g. histogram) are examine visually to
determine logical breaks (clustering) in the data.
6.
Optimal: The optimal
classification method is a solution to the limitations noted for maximum and
natural breaks. The optimal method places like data values in the same class by
minimizing and objective measure of classification error (Table 4.5, page 70).
Generally
speaking classification methods that consider how data are distributed along
the number line (such as natural breaks and optimal) are desirable because they
place similar data values in the same class (and dissimilar values in different
classes). Other methods such as equal interval and quantiles, however, may also
be desirable because they satisfy other criteria e.g. no gaps and easy calculation. If the purpose of
classification is to simplify the appearance of the map, spatial context should
also be considered (values of enumeration units as a function of surrounding
units, page 79).
