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).
Note
: Creating a grouping followed by executing a constrain (spatial or by attributes), will produce a different map then if the constrain is first applied and then the grouping is performed.

This module presents the three main grouping techniques described by Dent [1985]:

    • Equal step;
    • Quantile ;
    • Statistical ; and
    • Unique value.

 Grouping by the Equal step technique:

This grouping technique generates n groups having the same range of values in each group. Its computation is performed as follows:

       Range = Vmax - Vmin / n

Where:
Range = range of values in each group.
Vmax = attribute's maximum value.
Vmin = attribute's minimum value.
n = number of groups.

Example: Suppose one is studying soybean farms and has production data, number of employees, fertilization amount, etc. It is possible to group these farms by any of these attributes. Select, for instance, the soybean production, the first step is to compute the range for each group. For this it is required to analyze the soybean production data and extract the maximum and minimum values. Next one should define the number of groups that has to be generated.

Now, suppose the following information is available:
Number of groups = 3, minimum production = 3000 and maximum production = 93000.
Range = 93000 - 3000 / 3;
Range = 30000;

Then, we have:
Group 1: 3000 to 33000
Group 2: 33000 to 63000
Group 3: 63000 to 93000

In this way, farms producing from 3000 to 33000 will be classified as low production. Farms producing from 33000 to 63000 will be classified as medium production, and farms producing above 63000 will be classified as high production.
The important issue in this method is to notice that the range in each group slot is the same, in this case 30000.

 Grouping by quantile:

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.

       Nobjs = NTotal / n

Where:
Nobjs = number of objects in each group.
NTotal = total number of objects.
n = number of groups.

The range for each group is the interval found among objects in the sorting array.

       sort(n-1) <= Range(n) < sort(n)

Where:
Range(n) = range for the n group.
sort(n-1) = attribute value for the element in the n-1 position in the sorting array.
sort(n) = attribute value for the element in the n position 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.
In this way, if n = 4, will result in 4 groups, each having 25% of the farms. If n = 5, will result in 5 groups, each having 20% of the farms.

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).
The number of groups generated above and below will depend on the selected standard deviation value. Thus, the number of groups depends on this value and also on the existing objects attribute values.

Example using a 1sd interval:

 

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.

As an example consider the case where farms will be grouped such that group1 will have all farms with mortgage, and group2 will have all farms without a mortgage.

group1: mortgage=yes
group2: mortgage=no
In this way it is possible to generate a map showing the farms with a mortgage.

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.

Executing a grouping:

  1. The grouping module is activated from the interface " Visualizing Attributes" (inside a QUERY - Generating and Selecting Collections). Notice that this interface shows in its body a list with all objects categories existing in the information layer (IL). Normally it is recommended to place just one object category in each IL. However, when it is not the case, the user has to select one of the categories by double clicking on the desired category. This operation highlights the selected category for the grouping.
  2. press on Edit in the "Visualizing Attributes" window and select the Grouping... item. The dialog box "Grouping Objects" is presented.
  3. select among the Equal step, Quantile, Statistics orUnique value mode. Do not select the Pie Chart or Bar Chart.
  4. if the selection is the Equal step or Quantile, select the number of parts using the Number of Parts option. If the selection is Statistics, select the interval among groups through the Standard Deviation option.
  5. select an attribute in the Attribute's list.
  6. select "none" in the Normalize list. Later, in other tests, repeat the selecting procedure.
  7. select a color in Color Gradation;
  8. presson Group. The grouping interval and its respective colors are presented in the module (not over the map).
  9. presson Apply. The result is presented on the screen.
  10. it is possible to make other groups by selecting other options. Except the color, notice that the changes are effective when the Group option is pressed and, the final result is shown on the map, only when the Apply button is pressed.

Undoing a grouping:

  • to undo a grouping, select the mode in one of the 4 grouping modes, and next press on Ungroup. The module removes the groups temporarily. In order to make it effective press on Apply.

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.

Changing the Grouping Values and Titles Manually:

  1. press twice on a line representing geo-objects group (as in the figure below, where the "11-12" Income slot is selected), notice that the interval is presented in the text box below;
  2. change the limit values and press on CR; Notice that the module does not allow that the limits trespass the surrounding groups. When the values are accepted, after pressing CR, the new values replace the old ones. When the given values are invalid the replacement will not happen.
  3. In the second text box below, type a title to associate to the respective grouping and press on Associate.
  4. if desired change the values for other groups.
  5. DO NOT press on Grouping. This button, if pressed, will erase the values changed manually, that is, the system will recompute them. The Grouping button will restart the visual aspects and the values for each group.
  6. press on Apply so the changes will take place.

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.

Changing the Grouping Visual Aspects:

  1. press twice on a line representing a geo-objects group. The selected group will be highlighted.
  2. press on Visual. The dialog box "Graphic Presentation Views" is presented. This module allows to change the visual attributes.
  3. make the desired changes and, still in the "graphic Presentation Views" window press on Apply.
  4. in the grouping window , DO NOT press on Grouping. This button, if pressed, will erase the values changed manually, that is, the system will restart the visual aspects.
  5. in the grouping window, press on Apply so the changes will take place. on the map.

Changing the category visual aspects temporarily:

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.

  1. if there is a group selected (highlighted) press on it. The module removes the highlight.
  2. press on Visual.... The dialog box "graphic Presentation Views" is presented. This module allows changing the visual aspects attributes, in this case, changing temporarily the visual aspects presentation when the objects are ungrouped.
  3. make the desired changes.
  4. finish the changes and, in the "graphic Presentation Views" press on Apply.
  5. now , in the grouping window press on Apply so the changes will take place on the map.

NOTICE: This visual aspect is kept until the user exits the system.

Temporarily Changing the Visual Aspects of Polygons Outline:

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.

  1. apply any type of grouping (follow the steps of the Executing a grouping topic). After the application, execute the following steps:
  2. if there is a group selected (highlighted) click on it. The module removes the highlight.
  3. click on Visual... The dialog box " Setting Visual Parameters" is presented. This module allows changing the visual aspects attributes, in this case, change the visual aspects for the lines presentation. Change , for example, color and thickness.
  4. click on Apply..
  5. in the grouping window, press on Apply so the changes will take place on the map.

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:

  1. select the Pie Chart mode.
  2. select an attribute in the list and then press the Insert. The module shows the selected attribute with an associated color.
  3. repeat step 2 to add other attributes. The module allows a maximum number of 10 attributes.
  4. select an item in the Dim. Proportional to list. The attribute value selected determines the drawing dimension for the pies. To start, it is recommended to select the "none" item so all pies will be drawn with the same size. Later, in other tests, select an attribute to change the pie dimension.
  5. select the minimum size, pressing on Minimum Size. This value determines the diameter, in millimeters, for the smallest pie that can be shown on the map.
  6. select the maximum size pressing on Maximum Size. This value determines the largest diameter, in mi limiters, for the largest pie that will be shown on the map.
  7. press on Apply. The result is presented in the screen.

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:

  1. select the Bar Chart mode.
  2. select an attribute from the list and press on Insert. The module shows the selected attribute with an associated color.
  3. repeat the step 2 to insert other attributes.The module allows a maximum number of 10 attributes.
  4. select the bar width pressing on Width. This value sets the width, in mi limiters, for the bars that will be drawn on the map.
  5. select the minimum size pressing on Minimum Size. This value determines the height, in millimeters, for the smallest bar shown on the map.
  6. select the maximum size pressing on Maximum Size. This value determines the height, in millimeters, for the largest bar that will be shown on the map.
  7. press on Apply. The result is presented in the screen.

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.

See also:

Database Query
How to control Objects Visualization .
Attributes Query.
Spatial Query
Attributes Table.
Pointing and Photo Module.

Query Models .