I think your first point is valid, though clearly numbers-crunching has been shown to have value in non-quantitative contexts. Look at baseball.I mostly think this is horseshit, despite it arriving at some mostly believable conclusions. The first reason is because it's a classic case of a bunch of STEM folk deciding to help the poor music academics "scienctificize" their parlor games and then proceed to tell the music folk "facts" about their own field. Why were there no musicologists in this study? The second reason is that it used pop music charts and other popular consumption-based charts for their data pool, while ignoring the fact that these lists didn't just happen in a perfect system where people seek out music and just make it popular. These lists were carefully curated and implemented by big money like Clear Channel and record/media companies. Sure, popular tastes have some effect on what becomes popular and what trends, but for the past 15-ish years at least, what becomes popular has been mostly engineered, and you could even argue that this engineering creates popular tastes.
I don't think that matters so much, though - they're not trying to explain why patterns in changing tastes happen, so much as document periods in which they did change. I agree with you that the increasing concentration of the music distribution industry and the greater power a smaller number of individuals has to influence popular taste (or, maybe more precisely, more carefully deliver music which will appeal to the greatest percentage of current taste, which probably impacts the ability of music to evolve) has absolutely changed how music is consumed, no arguments.I guess my point is that this study claims to chart popular tastes as a somewhat independent metric, without accounting for or acknowledging the massive confounding variable that is active manipulation of popular tastes for commercial reasons.