Hi this is my first post here. I agree that Behe often relies heavily on lab experiments, but I think that he is doing that because he is asking a very different sort of question than most evolutionary biologists are. They are asking how evolutionary mechanisms evolved biological functions, while Behe is asking are evolutionary mechanisms alone capable of creating all biological functions. So I think that he wants to use examples where evolution is demonstrably the cause of a given biological change, as opposed to inferring it as the cause of an unobserved event.
I always thought that the mathematical evidence is more interesting, because, I think that either evolutionary theory is demonstrating something fundamental about algorithms, or it isn’t actually a realistic approximation of reality. Since you are a data scientist you might like this paper below which reviews the performance of evolutionary algorithms:
The limitations in the ability evolutionary algorithms of evolutionary algorithms, is precisely what an ID theorist such as Dempski would expect it to be, and precisely contradicts the expectations of the theorists who originally designed the evolutionary algorithms in the first place, so the question is why? If ID is actually a pseudo science as is often claimed then why does it accurately describe why evolutionary algorithms, are self limiting? Might better computer models of cells developed in the future similarly demonstrate mathematical limits to biological evolution consistent with ID someday as well? Perhaps ID is just a young science, with many imperfect personalities who are making many mistakes, but it’s fundamental premise is closer to how the world actually works than the one presented by evolutionary theory. That’s how I see it anyway.