![]() Grouping Module
The Grouping
module controls the way to symbolize geographic objects based on its
characteristics attribute values. When activated for the
first time this module recovers the state from the last section. Then, when
pressing on Apply it is possible to get a map with the
recovered views. However, this may not reflect the same results presented in
the last section. This will happen only if the groups defined in the last
section were computed after applying one constrain (spatial or by
attributes). This module presents the
three main grouping techniques described by Dent [1985]:
This grouping technique
generates n groups having the same range of values in each group. Its
computation is performed as follows: Where: Now, suppose the
following information is available:
This group technique
generates n groups having the same number of objects in each group. This is a
way to separate groups in a percentage way. The values range for each group
is computed by sorting the objects by the selected attribute. After the
sorting step, the number of objects in each group is computed, which is given
by the total number of objects divided by the desired number of groups.
The range for each group
is the interval found among objects in the sorting array.
Following the soybean
farms example, the range values for each group will be different, but each
group will have the same number of farms. Then, grouping by quantile, using n = 3, will give 3 groups, where 33.3% of
the farms with the worst production will be placed in group 1, other 33.3% of
the farms with medium production will be in group 2, and the last 33.3% of
the farms with high production will be in group 3. Grouping by standard
deviation
This grouping technique generates groups separated by the standard
deviation value. In this grouping technique the main divisor is the mean µ. Starting
from the average the grouping is performed by selecting elements that are
above the mean or below the mean in intervals related to the standard
deviation, like one standard deviation (1sd), or half standard deviation
(0.5sd), or even a quarter standard deviation (0.25sd). Example using a 0.5sd interval: Grouping by a Unique value
This grouping technique
generates groups having distinct values. Usually it is used to separate
objects having well defined characteristics such as, for instance, dry or wet,
warm or cold, exempt or not exempt, debtor or not debtor, etc. This is a
grouping method where the range is not computed. It depends only of the
attribute value. The number of groups generated depends on the differences
among existing values. For grouping made by Equal step and Quantile, the option
Number of parts, allow grouping them in up to 15 parts. For the Statistics
mode, the option Standard Deviation, allows to select intervals, among
groups, of 1, 1/2 or 1/4 of the standard deviation. For the grouping by Unique value there is no option to divide into number of parts or intervals. The
system is in charge to generate the groups automatically as a function of the
values to be analyzed. The grouping module
interface allows to work the groups visual aspects (colors, patterns and
shading), represented by points, lines or polygons in the map. The system
defines automatically a Color Gradation, for each group, and keep other
visual aspects attributes, such that patterns, shadings, styles, types,
thickness and heights, unchanged. If the user has to change the visual
aspects, this has to be done manually. This module allows to test several ways to perform grouping before processing
it. In this way, to make a grouping and show the result in the screen, just
press on the Apply button. The quantity of different
colors, automatically generated by the system, depends on the selection of
the number of colors used for the scale. The module allows the selection from
a simple color up to a set of 4 different colors. This selection, in the
module, is performed by the Color Gradation button. When it is chosen
to make a scale using only one color, the quantity of colors generated
depends on the number of existing variations for this color. For instance,
for the cyan color there is a maximum number of 4
variations which allow to make a distinction of up to 4 groups. When the user
asks for a larger quantity of colors, the user has to select a color that
allows more variations or select a set with more colors. When the number of
groups is greater than the possible number of scales, the system repeats the
last color generated for the remaining groups. When this happens, it is said
that a color saturation happened.
NOTE: The captions
generated in the grouping is presented also in the visualization
control module interface, that is, in the dialog box "Visualizing
Attributes". Thus every time the user press the Apply
button in the "Grouping Objects" window, the new caption is
automatically passed for the visualization control module. NOTE: one should not select the Pie
Chart or the Bar Chart to make this operation.
ATTENTION: The user possibility to change
grouping values can create also groups with limit values that exclude some
geographic objects. NOTE: A manual change in the superior
limit value in the third group of the figure above exclude all districts
where the population has an average income equals 13 minimum wages. Then, all
districts, where the population average income equals 13 minimum wages are
not classified in any group in this module, so these objects are represented,
in the map, using only their contour.
When applying an
ungrouping, the objects are presented on the map with the same visual aspects
of the category. The grouping module allows a temporary change of the
presentation visual aspects. The procedure for change is similar to the one
used in the previous topic. The difference is that no group has to be
highlighted over the interface. To illustrate how this is done follow the
next steps, but, don't forget to ungroup before.
NOTICE: This visual aspect is kept until
the user exits the system.
The visual aspects of the
polygon outline on a map can be changed, temporarily. By default this visual
aspect is the same of the category. For this, follow the same steps of the
last topic and change the lines visualization attributes. As an example see
the following steps, but first, make any grouping.
NOTE: This visual aspect is kept until
the user exits the system. Showing Charts on a Map
The grouping module
allows to show charts in each geo-object represented
on the map. The charts can be either pie or bar. The charts assume an
attribute selection expressed in comparable units so there is a coherent
interpretation. The pie chart
shows an attribute proportion related to its summation. It allows to select an attribute giving the proportionality
dimension, allowing a comparison among geo-objects. For its dimension
computation the average of the maximum and minimum sizes is used. In the case
where no attributes is selected as a proportional dimension, all pies will
have the same diameter for all geo-objects. It is possible, for
instance, to select the male and female population attributes for each
district (geo-object) as attributes for the pie chart. In this way, each
district will have a pie chart showing the proportion of males and females
represented by colors associated to each attribute. It is possible to select
another variable, for instance family income, for proportional dimension. Thus,
each pie diameter in each district will be proportional to the family income
in the district related to the maximum and minimum family income values in
the district set considered. The bar chart
shows the proportion for each selected attribute related to the total
observed for each geo-object through the height for each bar, using the corresponding
color. The height for each bar is proportional to the attribute dimension and
this proportion is kept for all geo-objects in the considered set. In this
way, it is possible to compare directly in the chart if a certain attribute
is greater or smaller for the evaluated geo-objects. NOTE 1: in previous versions (up to
Spring 3.3) the chart presentation was performed exclusively related to
groupings. That is, to show a chart the system automatically applied an
ungrouping. But, starting at version 3.4 it is possible to show charts
simultaneously with any grouping. NOTE 2: this characteristic, when the user press on
ungroup maybe a bit confusing. The
button ungroup has two functions. When the grouping was performed by
Equal step, quantile, statistics, or Unique value
mode it clears the grouping presented in the interface. Do not forget that
the ungrouping only takes effect on the map when the Apply
button is pressed. In the pie chart or bar chart mode the ungroup button just
clears the selected attributes for exhibition. Again, do not forget, the ungroup
only takes effect when the Apply button is pressed. Showing a pie chart:
NOTE: When the "none" item is
selected in the Proportional Dim.
a list, all
pies are drawn with the same size. NOTICE 1: The module allows, also, to select
and to remove graphic attributes through the Remove button. NOTICE 2: To change the color associated to
an attribute repeat the same procedure from the Changing the grouping visual
aspects topic or, just select an attribute and press the Remove button
and next the Insert button. The system associates a new color to an
attribute automatically. Showing bar charts:
NOTICE: To repeat or change the color
associated to an attribute, follow the same
instructions of the previous topic. ATTENTION: It is important to remember that
all operations described here will only take effect when the user press on the Apply button. Save and Load a Grouping
Any grouping can be shown
if it was stored in the hard drive, so a name can be associated for the
grouping. Press on Save...
to activate a new interface. In this interface, the user has to type a
name to save the grouping and next press the Apply button. Press on Load... to activate a new interface. In this
interface, the user select a grouping name saved
previously, and then, press on the Apply button. The selected
grouping is loaded in the Grouping module.
Database
Query ![]() |