The numbers here are "and mathematical scientists"¹, so it's not (as upthread implies) CS grads, it's CS grads + other studies. I lived with a physics major; she did some code, would she have been counted?², but she was not a SWE in the making³.
My gut would not think other math degrees would necessarily be more women heavy. But like upthread, my class was 2 women in a class of ~160, or <2%. Around that time I recall seeing a Stanford T-shirt with their ratio at something like 1:16, and they're prestigious enough one would expect their ratio to be above average.
Not sure what to say, aside from I cannot reconcile it with experience, and the numbers being used here aren't the ones we need.
(¹as this is how the source for the data, "National Survey of College Graduates", has uselessly lumped them together. The "mathematical" portion includes degrees such as statistics, "Mathematics, general", and other unspecified-by-the-methodology degrees. Even the "computer sciences" portion isn't just CS degrees, they've also lumped, e.g., IT in there.)
(²no, probably not; when I wrote that I was looking at a graph another poster posted, but that graph seems to munch the category names. Likely physics would be under "Physical and related scientists", but also I can't find the methodology of then what a "mathematical science" is…)
(³and in lecture, the in-lecture ratios changed rather dramatically once you got past the point of "other degrees require 2 courses of CS cross-training")