I saw it suggested on Twitter yesterday that the political polling firms should be doing some coronavirus polling, tracking trends in self-reported cases the same way they track public opinion.
I'm not sure if that would be scientifically useful in any way, but it sure would be interesting.