Hello,
Thanks for your question! The selection of clustering type depends on the dataset you have and the objectives you want to fulfill. If you want to group variables in the dataset, you can use hierarchical clustering (agglomerative) methods. If you want to group observations in the dataset, it is better to use K-means clustering.

The K-means clustering works well with larger datasets. Hierarchical clustering (agglomerative) methods work well with both larger and smaller datasets.

If you use K-means clustering, you should specify (know) the number of clusters before training the algorithm. If you know the number of clusters, go for K-means. Otherwise, go for a hierarchical clustering (agglomerative) method.

Regards,
Rukshan.