First of all, you would agree that you can aggregate clusters of people based on how each answer a variety of probing questions, right?
Nope. I've taken that test a few times, got different results each time, same for a number of friends and colleagues. It's too vague, swings with mood and interpretation, and is wildly swayed by it's own popularity. It's about as accurate as a horoscope, and has as much to do with reality as a Hogwarts house (which, even in-universe, wasn't a reliable predictor of the character of a person).
Since your thesis is flawed, I didn't bother to do more than skim that wall of text, but what I saw also read like pseudo-scientific nonsense. You mentioned something about it "not being harmful"; Tell that to the people who - no shit - didn't get jobs in management or analytics because the sorting hat didn't like their 4-letter password.