(Too long for a comment and it is kind of a soft question anyways)
I'm not so sure that your assumption of an imbalance is valid. Maybe it is. But would be better if demonstrated (or at least explored) first. Otherwise, we end up finding an explanation for a phenomenon that doesn't actually exist. Plus the exploration would probably inform the answer, even if assumption shown correct. (For example perhaps there are differences in schools with/without grad programs.)
I looked at the course curriculum for USNA math and it seemed like there were about equal classes in algebra and analysis (maybe even more in algebra).
There were two semesters of abstract algebra and three of linear algebra. (Different difficulty "tracks" also, but I'm just talking the sequence.) The formal analysis only included two semesters of real analysis and one of complex analysis. Again with a confusing multiplicity of tracks, but just the three real semester courses. No functional analysis as the institution is undergrad only.
I guess you could call the normal computational normal calculus courses, "analysis". Or ODEs, PDEs. But I'm not so sure this is valid. There's definitely a different feel for what people think of or use as calculus (solving problems) versus real analysis (theory).
In addition, if you're going to count the foundational conventional calculus semesters, then I can always bring in the four (4!) semesters of high school algebra that students typically take in the US of A.
On the actual explanation, I would say that "calculus" (and diffyQs is calculus on steroids) is useful in science and engineering. But this is a different thing than the highfalutin math major "analysis" of your question. So I think there are more likely other drivers for the math major interest in the topicm, other than physics applications. Complex analysis, has way less application than calc 1-3 and diffyscrews. (Pedants: I'm not saying there is never a single application, but not enough to make the topic a general STEM need and a required course.) Real analysis even more so not needed, not required to be taken, etc. etc. So don't blame your functional analysis courses on science and engineering.
Note also, that HIGH SCHOOL algebra is a HUGE useful topic in science and engineering...freshman chem is essentially one giant algebra word problem. (In a similar manner to how calculus and diffyQs are work horses.) But did you even think about that when discussing algebra in the context of your question?