I'm watching a webinar by
https://www.anl.gov/profile/charles-m-macal, whose group is making the models that the City is using for guidance. They have a sophisticated agent-based model that models all households in Chicago and seems to fit the current data pretty well.
Here's an interesting prediction:
The y-axis here is a little confusing (I'm not sure if it's something like R0, related to the derivative of the prevalence or what. Otherwise the peak at 4/1 doesn't make sense to me).