This is an interesting observation, and ties neatly with my skepticism about “data science
as currently utilized. The rate at which data is generated is not unbiased: some phenomena (and people) generate data at far greater rate than others. Data about and for the wealthy and connected are generated much faster than those about the poorer and the marginalized. If we want to use data to enhance equality, then relying too much on data, or at least, too naively, is exactly the wrong thing to do. Rather than ameliorate inequality, it exacerbates it. It creates an echo chamber where not the interests of the poor and the marginalized are drowned out by the preconceived imaginations about them by the rich and the connected.
In a sense, this is partly the background behind the Michigan surprise–but it is on a small scale. How much are we getting wrong about the world because of this arrogant blindspot about how data is generated? Good question.