Agreed. Another example is I'm following some people not because they share the same interests as me, but because my interest in Steem at this time is more to study the dynamics of the system so I am observing some obscure example users to study their patterns of using the system.

It is very difficult to automagically measure groups, because as @complexring pointed out about the Grassmannian, everything is relative to everything.

P.S. see my reply to your other comment. Unfortunately Steem has a feature bug in that it doesn't notify of replies when the comment nesting is too many levels deep.