Well this is an…interesting way to display this data.

I have mixed feelings about the use of Chernoff faces (or any derivatives) to visually display multivariate data. Unless done correctly, I think that this type of visualization can be difficult to interpret. I also think it lacks one of the main purposes of a visual – a quick and relatively easy way to understand the data – because it involves a lot of parsing of the individual pieces (eyebrows, eyes, nose, mouth, face shape, etc.) which is a pretty big ask if someone’s just quickly glancing at the faces.

But anyway, I do like the idea of trying to make “Chernoff emojis” for each state based on the variables in question. Using the color of the faces as a representation of one of the variables rather than face shape/size makes them all look a bit more uniform and makes comparing the other features a bit easier.

Also:

Yeah, that’s my expression whenever I think of West Virginia, too.