Seconded. In my experience the data profile tables are often the natural starting point for working with census data. It would be awesome if they were available through the API.

This is a good idea, and something we've thought about, generally. The problem is that the Census bureau does not provide those tables as bulk data. We could reverse engineer the computations, but then we need to develop a system for having "pseudo-tables" mixed in with official Census Bureau tables.

I think this is something which could be done, but it may have to wait until we secure more funding. It's probably just a little more involved than we can pull off in volunteer mode.