Use K-Means clustering to cluster whisky distilleries by their tasting profile. Use the elbow or silhouette method to find the optimal number of clusters.
To see how successful clustering was, report relevant metrics (e.g. silhouette, adjusted rand index, etc.) and create a plot showing the different distilleries, their classes according to the k-Means clustering, and the distance between points. You can use sklearn.manifold to get Euclidean distances between points.
Describe the main differences between the cluster - what are the factors that differ between classes?