Cluster distribution includes Hot Spot Analysis (Getis-Ord Gi*) and Cluster and Outlier Analysis (Anselin Local Moran's I). The mapping clusters tools can perform cluster analysis to identify the locations of statistically significant hot spots, cold spots, spatial outliers, and similar features. For example, in crime analysis, we can study which locations have frequent and clustered criminal activities, which plays an important auxiliary role in police force deployment.
Cluster distribution is a statistical method for classifying research objects in data. Such clustering methods share a common characteristic: the number and structure of categories are unknown beforehand. The analysis is based on data measuring similarity and dissimilarity between objects. These similarity/dissimilarity measurements can be regarded as metrics of "distance" between objects. Objects with close distances are grouped into the same category, while objects in different categories are farther apart. This represents a common approach in cluster analysis methods.
Cluster analysis can address the following questions:
- Where do clusters or cold spots and hot spots appear?
- Where are spatial outliers located?
- Which features are highly similar?
Related topics
Measuring Geographic Distributions