Lab 5: Pattern Analysis

Spatial Statistics Tools for Pattern Analysis

This exercise involves using ArcGIS Spatial Statistics tools to analyze the spatial patterns of various datasets. Follow the instructions below and answer all questions.

Instructions

Download and save the lab files (Lab5.zip) to your workspace. You need to unzip the file to a folder on your drive. Follow the instructions below and answer all questions at the end of this exercise. Your maps may be in color for this lab.

The statistical tools used in the lab generate graphical results and geographical outputs. You will be expected to include the graphical results in your lab write-up. These will take the form of charts and tables. See below in data preparation for where to access this data.

Data preparation

* Start ArcMap and click on the "Catalog" tab on the right-hand side of the window. Create a personal (or file) geodatabase called "lab5" on your drive.

* Open ArcToolbox (if it is not open already) by clicking the icon at the top.

* Click on the Geoprocessing menu and select Results. This will add a tab to your ArcToolbox window (at the bottom). The reports and graphics from tools you run will be accessible here for you to copy and paste into your lab report. These will typically show up as an HTML in the results window. Double-click on it and it will show in your internet browser. DO NOT INCLUDE THESE ON YOUR MAPS.

Exercise 1: Moran's I and Conceptualization of Distance

1. Import the Crime shapefile into your geodatabase. Name it "Crime".
2. Click the Add Data button and add Crime feature class to your map.
3. Change the symbology to show a choropleth map of robberies (ROBB01 in the attribute table) using 7 Natural Breaks classes. (Do not use a diverging color ramp)
4. Open the Spatial Autocorrelation (Moran's I) tool under Spatial Statistics > Analyzing Patterns.
5. Set the input feature class to Crime and the input field to ROBB01.
6. Check the box next to Generate Reports. This will add graphical outputs to your results window in the form of an HTML.
7. Confirm that the conceptualization of distance is set to Inverse Distance and enter 150000 for the Distance Band. Leave all other fields blank and click OK.
8. Look at the graphical results for this tool.
9. Open the Cluster and Outlier Analysis (Anselin Local Moran's I) tool under Spatial Statistics Tools > Mapping Clusters.
10. Set the input feature class to Crime and the input field to ROBB01. Save the output to your geodatabase.
11. Confirm that the conceptualization of distance is set to Inverse Distance and enter 150000 for the Distance Band. Leave all other fields blank and click OK.
12. Once the output has been added to your map, change the symbology to show a choropleth map of the LMi Index using the Hot to Cold Diverging color ramp with 7 natural breaks classes.
13. Repeat Steps 4 - 12 with inverse distance squared as the conceptualization of distance. Leave all other fields the same.
14. Repeat Steps 4 - 12 with polygon contiguity (first order) as the conceptualization of distance. Leave all other fields the same.
15. Generate ONE map with the choropleth display of the robbery data and all three map results from the Local Moran's I tool. Include a text box on the map explaining how the conceptualization of distance impacts the results of the statistics.
16. Include separately the 3 graphical outputs (charts and tables) from the Spatial Autocorrelation (Moran's I) tool.

Exercise 2: Getis-Ord

17. Import the Mortality shapefile into your geodatabase. Name it "Mortality".
18. Click the Add Data button and add the Mortality feature class to your map.
19. Change the symbology to show a choropleth map of unintentional injuries, white females (RATEWF10 in the attribute table) using 7 classes and Natural Breaks. (Do not use a diverging color ramp)
20. Open the High/Low Clustering (Getis-Ord General G) tool under Spatial Statistics > Analyzing Patterns.
21. Set the input feature class to Mortality and the input field to RATEWF10.
22. Check the box next to Generate Reports. This will add graphical outputs to your results window in the form of an HTML.
23. Confirm that the conceptualization of distance is set to Inverse Distance and enter 150000 for the Distance Band. Leave all other fields blank and click OK.
24. Look at the graphical results for this tool.
25. Open the Hot Spot Analysis (Getis-Ord Gi*) tool under Spatial Statistics Tools > Mapping Clusters.
26. Set the input feature class to Mortality and the input field to RATEWF10. Save the output to your geodatabase.
27. Confirm that the conceptualization of distance is set to Inverse Distance and enter 150000 for the Distance Band. Leave all other fields blank and click OK.
28. Once the output has been added to your map, change the symbology to show a choropleth map of the GiZScore using the Hot to Cold Diverging color ramp with 7 natural breaks classes.
29. Generate ONE map with the choropleth display of the white female, unintentional injury data and the map result from the Getis-Ord Gi* tool.
30. Include separately the graphical outputs (charts and tables) from the High/Low Clustering (Getis-Ord General G) tool.
31. Complete the following short essay:  In your own words, describe the results, map, and why you think the distribution is as it is.

Exercise 3: Ripley's K and Distance Bands

32. Import the Ag shapefile into your geodatabase. Name it "Ag".
33. Click the Add Data button and add Ag feature class to your map.
34. Change the symbology to show a choropleth map of Acres of Harvested Corn (M163_07 in the attribute table) using 7 Natural Breaks classes. (Do not use a diverging color ramp)
35. Open the Multi-Distance Spatial Cluster Analysis (Ripley's K Function) tool under Spatial Statistics > Analyzing Patterns.
36. Set the input feature class to Ag and save the output table to your geodatabase.
37. The number of distance bands should be set to 10.
38. Change the number of permutations to 9. (The tool will take a while to run, setting it to any larger value will cause to the tool to take a REALLY LONG time to run. Please do not set it about 9.)
39. Check the box to Display Results Graphically
40. Set the Weight Field to M163_07 and leave all the other fields. Click OK.
41. The graphical results will pop up automatically.  DO NOT CLOSE THIS WINDOW UNTIL YOU HAVE COPIED IT!  Right-click on the graphic and select "Copy as Graphic". Then go to your word document and paste it in. Include a title. Once you close this window you will have to rerun the tool to get the image again.
42. Repeat Steps 35 - 41 with the Beginning Distance set to 50 and the Distance Increment set to 5000. Leave all other fields the same. Copy the graphical results to your lab write-up. Include a title to differentiate this result from the others.
43. Repeat Steps 35 - 41 with the Beginning Distance set to 50 and the Distance Increment set to 10000. Leave all other fields the same. Copy the graphical results to your lab write-up. Include a title to differentiate this result from the others.
44. Repeat Steps 35 - 41 with the Beginning Distance set to 50 and the Distance Increment set to 20000. Leave all other fields the same. Copy the graphical results to your lab write-up. Include a title to differentiate this result from the others.
45. Complete the following short essay:  In your own words, explain how distance impacts the results of running this statistic.

Exercise 4: Choose Your Own Statistical Adventure?!

46. Select a different field from either Crime or Mortality. Here is a sample of available fields.

Crime: Murd01 = murders, Assa01 = aggravated assaults, Burg01 = burglaries, larc01 = larcenies, mv_th01 = motor vehicle thefts, Arso01 = arsons

Mortality: wm = White males, wf = White females, bm = Black males, bf = Black females;  01 = heart disease, 04 = cancers, 10 = unintentional injuries, 15 = suicides

47. Change the symbology of your chosen data set to view the field as a choropleth using 7 natural breaks classes.
48. Run all the statistical tools, set to the parameters listed, on your chosen field.
1. Spatial Autocorrelation (Moran's I): Check Generate Report, Inverse Distance, and 150000 Distance Band
2. Cluster and Outlier Analysis (Anselin Local Moran's I): Inverse Distance, and 150000 Distance Band
3. High/Low Clustering (Getis-Ord General G): Check Generate Report, Inverse Distance, and 150000 Distance Band
4. Hot Spot Analysis (Getis-Ord Gi*): Inverse Distance, and 150000 Distance Band
5. Multi-Distance Spatial Cluster Analysis (Ripley's K Function): 10 Distance Bands, 9 permutations, and Check Display Results Graphically.

Leave all other fields with their defaults.

49. Generate 3 maps.
1. The choropleth map of your chosen field, symbolized with 7 Natural Breaks classes.
2. Show the graphical output from the Moran's I and the Local Moran's I map side-by-side.
3. Show the graphical output from the Getis-Ord General G and the Getis-Ord Gi* map side-by-side.
50. Include separately the Ripley's K graphical results.
51. Complete the following short essay:  What can you tell from the statistics that you can't from the choropleth map?