Gonna agree about the new popularity being kind of a pain to use - I've been building a website using this data over the past few weeks but the newest batch is a bit unreadable - could we get a developer to give some insight into the changes?
Why does an artax riven (when there was clearly only one traded this week due to med = mean & std dev. of 0) Have the same popularity as a broken scepter or spectra riven (which I estimated to have around at least 20 trades last week).
If there's a reason behind the change, I'd love to hear it.
Of course, I'm also for a full release of the data pre-parse (without usernames) but I'm also not a security expert or anything, just a nerd with free time.