It seems really hard to justify treating those 8 characteristics as independent, equally important, and all-encompassing variables. The page's attempt to do so

It seems really hard to justify treating those 8 characteristics as independent, equally important, and all-encompassing variables. The page's attempt to do so amounts to saying it's a peer-reviewed method employed by other music researchers. But the paper they linked made an effort to address that concern, and weights the features; it includes more echo-nest features, like tempo, mode and time-signature; it uses a different metric; and while it's not the overall conclusion of the paper, it includes a similar measure of similarity over time, but pegs 2015 as the most diverse year since 1963 in general, and 2014 as the most diverse year ever for #1s.