6.1 Topics Covered

• Understand the principles behind K-means clustering
• Know the requirements to carry out K-means clustering
• Combine dimension reduction and cluster analysis
• Interpret the characteristics of a cluster analysis
• Assess the spatial representation of the clusters
• Carry out a sensitivity analysis to various parameters
• Impose a bound on the clustering solutions
• Use an elbow plot to pick the best \(k\)
• Use the cluster categories as a variable

GeoDa Functions

• Clusters > K Means
  • select variables
  • select K-means starting algorithms
  • select standardization methods
  • K-means characteristics
  • mapping the clusters
  • changing the cluster labels
  • saving the cluster classification
  • setting a minimum bound
• Explore > Conditional Plot > Box Plot
• Table > Aggregate
• Tools > Dissolve

Toolbar Icons

Figure 6.1: Clusters > K Means | K Medians | K Medoids | Spectral | Hierarchical