Technical Notes: Hot Spot Analysis (Gettis-OrdGi*)

Introduction: Spatial Statistics assess patterns, trends, distribution and relationships. We’ll use Spatial Autocorrelation tool to see if features are clustered or dispersed. The spatial Autocorrelation is based on Tobler’s First Law of Geography which says everything is related to everything-but nearby things are more related than things are far away. It’s the correlation of a variable with itself through space. The output gives Z scores for distance.

Hotspot analysis uses vectors to identify the locations of statistically significant hot spots and cold spots in data. Hotspot Analysis finds features with similar attribute values spatially cluster together. Hotspot analysis Getis ord GI* statistic can delineate clusters of features with values significantly higher or lower than the overall study areas mean/average. It calculates Z score and P value for each feature. A high Z score and small P value for a feature indicates a significant hot spot. A low negative Z score and small P value indicates a significant cold spot. The higher (or lower) the Z score, the more intense the clustering. A Z score near zero means no spatial clustering.

The example shows distribution of social characteristics (Female householder) and Hospital readmission penalties for selected hospitals in Illinois. The penalty data was collected from Kaiser Health (kaiserhealthnews.org) and Female householder data was downloaded from American Community Survey. I built a thematic map from data of Female householder by ZIP and then layered it with penalty data. It looks like there is some relation between hospitals with high penalty rate and location where higher concentration of female householder reside. We’ll use hotspot analysis of Female householder data to see if the clusters are statistically significant.