I agree that Marxism is a bad theory of history and of economics, but it simply isn't that case that Marx didn't rely on data in his work. His work (and, perhaps to an even greater extend, Engel's work) is chock-full of data that confirms his theory. This is actually one of the examples Popper uses to demonstrate the uselessness of data-theory fit as a demarcation criterion for science. Confirmation is simply too easy to come by (especially, when you are more or less avoiding places where disconfirming data is likely to show up).