I have given a public talk at the LA East R Users meetup on July 8, 2019, at the USC Health Campus. The slides on the meetup GitHub are here.
A three line summary of the presentation is as follows:
- Never work with anything that doesn’t pass HMC checks or wildly goes wrong with PPCs.
- Do rstanarm/brms first than lower Stan code + rstan which has good default values and built-in pathology handlings.
- Post things on the Stan Discourse Forum, and don’t be afraid to try Stan out.