Geography Reference
In-Depth Information
values with most values being somewhere in between? When values are grouped into
classes they can be depicted using a histogram. h is is a form of chart that has bars
of a size that is in proportion to the number of values in a given class. For example,
if there are i ve observations in class 1 and 10 observations in class 2 then the bar rep-
resenting class 2 will be twice as high as the bar representing class 1. Figure 3.1 shows
a histogram with the range (minimum and maximum) of values represented by each
bar indicated (e.g. values in the i rst bar range from 27.1 to 29). Frequency indicates
the number of cases in a bin or class. As an example, there are 10 cases in the range
37.1 to 39.
Using too few or too many classes will not enable representation of important fea-
tures of the distribution. h e number of classes is likely to be determined as a function
of the number of observations and the range of values that they take. If there are many
thousands of observations (as might be the case, for example, using a remotely sensed
image), and there is a sui ciently wide range of values, then it may be sensible to have
a large number of classes, producing a 'smoother' distribution than would be possible
for a smaller number of observations.
h e most common way of summarizing a data set is to compute some kind of aver-
age, the mean average is the most well known. Averages are measures of central ten-
dency in a distribution—in some sense the 'middle' value in the distribution. h e
mean average of the variable, and the notation used to represent this, is detailed below.
First, a value of the variable is indicated by
z
i
. h e value itself is given by
z
, and
i
is an
index—it indicates the observation number. For example, if there are i ve observations
20
10
0
27.1-29.0
31.1-33.0
35.1-37.0
39.1-41.0
43.1-45.0
33.1-35.0
29.1-31.0
37.1-39.0
41.1-43.0
45.1-47.0
Value
Figure 3.1
Histogram. Values are precipitation amounts in millimetres.
Search WWH ::
Custom Search