I have a question I need help with and maybe someone could assist.

I'm building a system for a class that predicts student attendance.

I have 112 Data points per class (1 attendance record per week for the last 2 years+) for a grouping of 50 classes, and I need to set up an alert and find a 2 numbers

A) I need to find the attendance number that falls below a certain statistical significant per week. This will be used to ignore those classes from proceeding to step 2. (IE: If i have 23 classes that are 200 people and one that is a class is only 3 people, i dont care what the variance is in that class.)


B) An alert when there is a statistically significant variance in the amount of actual attendance verses the predicted amount. (what is the variance number that would indicate the alert)

I have data that shows actual attendance and data that shows predicted attendance (I think he used simple regression) on a weekly basis for the past 2 year. (yeah some professors do that)

I really need help in figuring out how to best find these numbers no matter how complex the statistical method I need to use is. Thought I'd ask everyone else how they would come up with these numbers?


I'm going to initially do this in excel to prove the formulas.. that thought is for it to scale.

Thanks for your advice