Go back and look at the scatter plot. It's a weak, but real correlation. The random deltas between nations are well above the significance of the gender signal. There's good science to be argued about there.
But it's being used here to justify an outrageous outlier. Women aren't just "less interested" in sofware at the scale we see in that study, they're outnumbered by literally a whole order of magnitude. Nothing from that study argues for this kind of effect, nothing at all.
