Data explorer allows us to explore each class individually.
We can filter based on cluster id, displaying one or more of the classes by themselves, each class can then be split into further subgroups, explored individually again, or as required, we can merge them together and get the original class back.