Hierarchical classification builds a community tree instead of a fixed number of families. You can see from the tree at what level the groupings occur, and then decide where to cut.
Type allows you to categorize either the individuals, that is, the variables :
⇒ sort the individuals answers the question «Which pieces are alike?»; ;
⇒ sort the variables answers the question «Which characteristics change together?» and usefully complements the ACP.
Label (optional) selects the column that lists the leaves of the tree.
The height of a cluster measures the loss of homogeneity it entails. Cut precisely during a big vertical jump : The number of branches obtained in this way is the natural number of families.
