One example of an area of science becoming a harder science, as a result of data science, is genomics emerging from genetics. Now we have an expansive body of data from which to derive insights.
I think I'd say that hard science is defined by minimal ambiguity. In math, there are million dollar prizes reserved for those who can crack some of the outstanding ambiguities. Whereas, in psychology, to my knowledge, ambiguity is the norm.
We can extend the epistemological spectrum further into ambiguity yet, moving from hard science to soft science to the likes of alchemy and ultimately art.