Support finite population sizes

[mnick]

Right now, our model assumes an infinite/unbounded set of possible events. However, in reality each county has only a finite population size that can be infected. Right now this doesn't us much, but it will become relevant as the number of infections grow.

There is a simple fix to support this as discussed here: https://arxiv.org/abs/1711.01679 in Equation 7.

To support this in our models. We'll also need to collect data about the population size in each county.