How do you select the appropriate prior distribution for a Bayesian estimation of driver acceptance rates on Uber Drivers?