Visualize the hierarchical structure of your data.
Definition
Agglomerative hierarchical clustering (AGNES) builds a hierarchy of clusters represented by a dendrogram. Unlike K-means, it does not require specifying k in advance and allows exploring different levels of grouping.
When to use it
Explore the natural hierarchical structure of data
Visualize relationships between observations with a dendrogram
When the number of clusters is not known in advance
Genomic, ecological, or textual data
Requirements
Continuous variables (or pre-computed distance)
Moderate sample size (< 500 recommended for readability)
Different linkage methods available (Ward, Complete, Average)
What StatsLab computes
Interactive SVG dendrogram
Cluster assignment (cut at chosen height)
Descriptive statistics per cluster
Methods: Ward, Complete, Single, Average, Centroid
Worked example
Context : Hierarchical classification of 30 countries based on 5 development indicators.
Result : 4 distinct groups at cut h=8: Developed, Emerging, Developing, Fragile
Interpretation : The dendrogram reveals two major super-clusters: high-HDI vs low-HDI countries. Ward method produces balanced cluster sizes.