Lab 9

Overview

In this lab, you will evaluate the degree of spatial autocorrelation (SAC) for multiple variables at multiple spatial scales. It is recommended that you use GeoDa for this assignment. First, download one of the two Wisconsin hazards datasets:

Questions

  1. Create a simple map of each of the following variables: flod_evt (flood events), hail_evt (hail events), and torn_dth (tornado deaths). No legend is needed (3 points).

  2. Make a prediction about the presence/absence of significant SAC for each variable. Do you believe it exists? Why or why not (1.5 points)?

  3. Complete the following table using Moran’s I and 3, 6, and 9 nearest neighbors (9 points).

Variable knn Moran’s I p-value Conclusion (+/-/no SAC)
Flood events 3
Flood events 6
Flood events 9
Hail events 3
Hail events 6
Hail events 9
Tornado deaths 3
Tornado deaths 6
Tornado deaths 9

  1. Which variable has the greatest amount of SAC according to Moran’s I? Do you think this would be different if more neighbor schemes were tested? Did any conclusions change by adjusting the number of neighbors? If so, which (3 points)?

  2. Next, you will compute Moran’s Local Index of Spatial Autocorrelation (LISA) on flod_evt using a neighborhood scheme of your choice. What did you choose and why (1 point)?

  3. Create a significance map (showing p-values) of flod_evt’s LISA. Where is the most significant spatial autocorrelation present within the state (2 points)?