> This study captures P(running into a profession) x P(preference for that profession). I would love to them model P(running into a profession), so that we could get the much juicer preferences of each group for various professions.
The page links to a different visualization of the same data, [0] which attempts to control for the prevalence of a job within the population.
Maybe it doesn't go far enough for you but that's a pretty big first step in that model.
> so that we could get the much juicer preferences of each group for various professions.
You'd probably just get a bunch of data that tells you the more female a profession is the more it wants to marry up and the more male a profession is the more it wants to marry toward a profession with a lot of young women.
The page links to a different visualization of the same data, [0] which attempts to control for the prevalence of a job within the population.
Maybe it doesn't go far enough for you but that's a pretty big first step in that model.
[0] https://flowingdata.com/2017/08/28/occupation-matchmaker/