Getis-Ord Gi* Hot Spot Analysis

Getis-Ord Gi* hot spot analysis computes, for each location, a z score comparing the sum of values in its neighbourhood including itself with the sum expected under spatial randomness.

Getis-Ord Gi* 핫스팟 분석은 각 지점마다 자기 자신을 포함한 이웃 값의 합을 공간적 무작위에서 기대되는 합과 비교한 z 점수를 계산해 높은 값과 낮은 값이 모인 곳을 찾는 방법입니다.

Where are the statistically significant hot spots and cold spots of crime rates across these neighbourhoods?

동네별 범죄율에서 통계적으로 유의한 핫스팟과 콜드스팟이 어디인지 알고 싶어요.

Define neighbours with a spatial weights matrix, compute the Gi* z score for each location and adjust the resulting p-values for multiple testing such as with the false discovery rate. Significant positive z scores mark clusters of high values and significant negative z scores mark clusters of low values.

When to use Getis-Ord Gi* Hot Spot Analysis

• You have a numeric value at many point or area locations
• You want to map where high or low values cluster
• You need location-specific results rather than one global index

When to use something else

• You need to separate spatial outliers such as high values surrounded by low values; use local Moran's I (LISA)
• You only need to know whether clustering exists overall; use Moran's I or Geary's C

Assumptions and requirements

• The spatial weights reflect the scale of the process; neighbours come from the coordinate columns (8 nearest, binary, including the location itself, by default)
• Values are non-negative for the raw Gi* ratio, and the normal approximation needs enough neighbours
• Multiple testing across locations is controlled

What the analysis reports

• Gi* statistic and z score per location
• Two-sided p-value, FDR-adjusted q-value and permutation p-value
• Hot spot and cold spot counts
• Map coordinates for plotting

How to report it

Getis–Ord Gi* analysis with [weights definition] identified [value] hot spots and [value] cold spots among [value] locations at FDR-adjusted q < [value].

Related methods

• Local Indicators of Spatial Association (LISA)
• Moran's I Spatial Autocorrelation
• Geary's C Spatial Autocorrelation

Run Getis-Ord Gi* Hot Spot Analysis on Quorant