Advanced ESDA

Exercise 12 Advanced ESDA

12.1 Objectives

This exercise illustrates some more advanced visualization techniques in ESDA in the form of map animation and conditional maps.

At the end of the exercise, you should know how to: • create and control a map movie • create a conditional map • change conditioning categories in a conditional map

More detailed information on these operations can be found in the User’s Guide , pp. 40–41, and Release Notes, pp. 26–28, and 38–40.

12.2 Map Animation

We continue to use the BUENOSAIRES sample data set. If this is not in your current project, clear all windows and load the file buenosaires.shp with Key variable INDRANO. The simple form of map animation implemented in GeoDa consists of automatically moving through all observations for a given variable, from the lowest value to the highest value. The matching observations are shown on a base map, either one at a time (Single), or cumulatively (Cumulative).

Invoke this function by choosing Map > Map Movie > Cumulative from the menu, as in Figure 12.1 (p. 87), or by clicking the toolbar icon. This brings up the familiar variables setting dialog (see Figure 11.3 on p. 80).

Figure 12.1: Map movie function.

Figure 12.2: Map movie initial layout.

Select AL99PC (for the “Alianza” party election results) and click OK to bring up the initial map movie interface, shown in Figure 12.2.

Click on the Play button to start the movie. The polygons on the map will gradually be filled out in a pink shade, going from the lowest value to the highest. Note how the map movie is linked to all other graphs and maps in the project, such that the selection in the movie becomes the selection in all other windows. You can stop the movie at any time, by pressing the

Figure 12.3: Map movie for AL vote results – pause.

Figure 12.4: Map movie for AL vote results – stepwise. Pause button, as in Figure 12.3. Clicking on Reset will wipe out all selected

polygons and start over with a blank base map. You can affect the speed of the movie with the slider bar (Speed Control): positioning the slider bar more to the left will increase the speed.

Once the movie is paused, it can be advanced (or moved back) one

Figure 12.5: Conditional plot map option.

observation at a time, by using the >> (or, <<) key. This is illustrated in Figure 12.4 (p. 88) for the map at a later stage in the animation process.

The purpose of the map animation is to assess the extent to which similar values occur in similar locations. For example, in Figures 12.3 and 12.4, the low values for AL99PC systematically start in the north eastern precincts and move around along the periphery, leaving the higher values in the city core. This is very different from a random pattern, where the values would jump all over the map. An example of such a random pattern can be seen in the grid100s.shp sample data set. Check this out for any of the randomly permuted variables (the variables starting with ranz).

12.3 Conditional Maps

The conditional map is a special case of the conditional plots considered in Section 10.2 on p. 69. Start the conditional plot function as before (see Figure 10.1 on p. 70) and select the radio button next to Map View in the view type dialog, as in Figure 12.5. Click on OK to bring up the variable selection dialog, Figure 12.6 on p. 90.

Select EAST for the X Variable, and NORTH for the Y Variable. Take the variable of interest (Variable 1) as TURN99PC. As in the previous examples, this is an illustration of geographic conditioning according to the location of the precincts, grouped into 9 subregions. Any other two conditioning variables could be selected, as in the generic conditional plot example. The interval ranges for the conditioning variables are changed by moving the associated handle to the left or right.

Finally, click on OK to bring up the conditional map, shown in Figure 12.7 on p. 90. The map employs a continuous color ramp, going from blue-green at the low end to brown-red at the high end. The color ramp is shown at

Figure 12.6: Conditional map variable selection.

Figure 12.7: Conditional map for AL vote results.

the top of the graph, with the range for the variable TURN99PC indicated.

The purpose of conditioning is to assess the extent to which there is

a suggestion of systematic differences in the variable distribution among the subregions. The maps in Figure 12.7 seem to indicate that the higher turnout precincts are on the west side and the lower turnout precincts on the east side. In a more elaborate exploration, other variables would be investigated whose spatial distribution may show similar patterns, in order to begin to construct a set of hypotheses that eventually lead to regression specifications.

12.4 Practice

Consider the three election variables for Buenos Aires more closely (APR99PC, AL99PC , and TURN99PC) to assess the extent to which they show similar or contrasting geographic patterns. Alternatively, revisit the rosas2001.shp corn yield example, or any of the other sample data sets you may have considered in Exercise 11.