If 10% of developers are female, there's a potential for 20% of all developers to be in male-female marriages. The data shown here had just 12% which could represent 60% of female software developers marrying male software developers (staggeringly high but not mathematically impossible).

I think there's two problems with that premise: (1) a large slice of that marriage number could be male-male or other marriages. (2) the figure "10% of developers are female" is true only if you're looking at global numbers. It's higher in the U.S. where these marriage figures originate.