Healthy Klamath :: Links to Cluster and Hotspot analysis of chronic conditions GIS

Hotspot and Cluster Maps

Background

LINKS TO MAPS ARE BELOW

The links provided below are given in three categories:

Hot-Spot Analysis
Cluster and Outlier Analysis
Walkability vs. Incident Density

A little background is necessary in order to properly interpret these maps.

For both the Hot-Spot Analysis and Cluster/Outlier Analysis procedures the data are categorized in terms of three spatial areas:

The area defined by the census block group, hereafter referred to as ‘the feature’
The area defined by the immediate neighborhood surrounding the feature, but not the feature itself. Note that the feature’s neighborhood is thus defined by other nearby census block groups.
The entire study area

In the Hot-Spot analysis, the 99%, 95%, and 90% confidence levels are found for each feature whose disease incidence rates for its neighborhood are significantly higher or lower than that of the study area.

In the Cluster and Outlier analysis, the disease incidence rates of the area of the feature (i.e., the area of the census block group) are compared to that of their associated neighborhood. This results in being able to ask the following two questions:

Is the feature’s neighborhood significantly different from all other neighborhoods within the study area?
Is the feature significantly different from all other features within its neighborhood?

There are four possible outcomes.

Clusters:

HH – the disease incidence rate in the feature area is higher than that within its neighborhood and the disease incidence rate of its neighborhood is higher than that of other neighborhoods in the study area
LL – the disease incidence rate in the feature area is lower than that within its neighborhood and the disease incidence rate of its neighborhood is lower than that of other neighborhoods in the study area

Outliers:

HL – the disease incidence rate in the feature area is higher than that within its neighborhood and the disease incidence rate of its neighborhood is lower than that of other neighborhoods in the study area.
LH - the disease incidence rate in the feature area is lower than that within its neighborhood and the disease incidence rate of its neighborhood is higher than that of other neighborhoods in the study area.

Note that this analysis can’t be done in any other way other than in a spatial context. Even though you could sort the data in a tabular construct to find the ‘high’ or ‘low’ values, these would be the ‘global’ high/low values. This process enables you to find the ‘local’ high/low value, i.e. outliers.

The areas that show HL or LH results are almost always the most interesting.

Hot-Spot Analysis Web Map Links: