Georgia Tech has a handy website that estimates the probability of encountering someone positive with covid19 based on the size of a group and the prevalence in the county: https://covid19risk.biosci.gatech.edu/
I checked the numbers on a few nearby counties and it seems to use the # of positive tests for the last 14 days divided by the county population to get a background prevalence. It then multiplies that by 5 or 10 based on your selection (i.e. 5 or 10 positive cases for every 1 positive test) and calculates the probability of encountering at least 1 person in a group of 10, 15, etc people...
For example, in hot spot El Paso a group of 10 random people would have at least 1 person positive for covid19 89% of the time, assuming there are 10 positive cases for every 1 positive test: