Improve wait time calculations

[ja5087]

The current wait time is a 75% confidence interval with a hardcoded expected avgHelpTime = 10 (minutes) divided by the number of staff available. Empirically we've found that this causes the interval to fluctuate a little too much. To see this, try varying the number of online assistants. We should develop a better way of calculating this and perhaps estimating parameters based on data instead of magic numbers.

[stella-lu]

By "based on data instead of magic numbers", is this suggesting that we keep track of times taken to help students in the past and basing avgHelpTime off of that?

[kristyelee]

I can try to help out with this issue