That stat is basically meaningless on its own. It could mean anything from they've done an amazing job on engaging women, to they've bodged the numbers by unfairly discriminating against men, or anything in between.

Annoyingly they actually do have the data to answer which it is, because Pycon's review process has a first stage which is blind, and a second stage which isn't. So if they published how many talks get rejected at each stage, by year and vendor, then we could draw actual conclusions.

I couldn't find where they have published those numbers though so we can't draw any conclusions here.