Analysis of attendance and churn

[sbenthall]

I'm starting a discussion here about research techniques for tenure, attendance, and churn.

There's already been some work on this:

#445

But this is a topic with a lot of depth to it, especially once we start looking for statistical evidence of what causes churn.

I wonder if a more systematic way of tracking individuals and extrapolating from point data to durational data would be useful here. See also:

#367