https://rpubs.com/heatherleeleary/hotspot_getisOrd_tut

R Tutorial: Hotspot Analysis using Getis Ord Gi Heatherlee Leary 2023-04-27

• Introduction
  • Example Scenario
  • Spatial Autocorrelation
  • R Packages
• Explore Raw Data
• Check Empty Neighbor Sets
• Global G Test
• Local Gi Test
• Disclaimer and Acknowledgements
• Session Info

Introduction

This tutorial will walk you through the steps and R code to perform

a hotspot analysis using the Getis Ord Gi method.

Note: Getis Ord Gi and Gi* are the same thing.

Example Scenario

As an example, we will look at tree equity score data for the city of Tucson, Arizona, USA. A tree equity score (TES) is “a metric that helps cities assess how well they are delivering equitable tree canopy cover to all residents. It is derived from tree canopy cover, climate, demographic and socioeconomic data” (American Forests 2023). TES values range from 0 - 100, with a lower score indicating lower equity.

Imagine we want to assess spatial patterns of TES across Tucson. Specifically, we want to know whether there are statistically significant clusters of high and low TES values. This information can help the City of Tucson prioritize where resources are allocated for tree planting initiatives and other climate mitigation strategies.

Spatial Autocorrelation

Before moving on, it’s important to understand the concept of

spatial autocorrelation.

Spatial autocorrelation is a measure of how a set of spatial features or their associated values are related to each other in space. It describes the extent to which nearby features or values are similar (clustered) or dissimilar (dispersed).