I have a user acquisition model that projects DAUs, MAUs and Registrations over the next 6 months. For infrastructure planning purposes, I want to model projected concurrency during peak time given DAU/MAU data. Is there a reasonable way to make assumptions around this?