[Calcite]

I like how they’ve compared ML clusters to US cities. Fundamentally, it’s still a continuous distribution between the five variables with four “hot spots”.

>>>Revelle likened the types to the location and population of cities. More people live in New York, Chicago, Los Angeles and Houston than anywhere else in the country, but most of the country doesn’t live in any of those cities. And although you can easily lump someone in Newark into New York, a person in Pittsburgh is harder to classify because they are equally close to New York and Chicago.