Hi @bridget sorry for taking such a long time to reply. I really appreciate your courage to give predictors a shot, but based upon what you said, I actually don’t think that you need to use predictor to make this threshold. You could simply make a data element add it to a data set ‘threshold’ data set and import a 1 value for it for any weeks you want.
The predictor could be used to count the number of incidents the threshold is violated. To do this you’ll need to make the threshold a constant and make a predictor with the following generator “IF(threshold>=cases,1,0)” Here is an example:
This will count the number of times the threshold is violated. You need to make your output data element to count. . You also need to schedule this predictor to run in the scheduler application. Below is an example of this.
I suggest you make your relative start at least 7 previous days (This would be represented as -7) and your relative end 1 day in the future. For the frequency of the predictor job you’ll need to make sure it is schedule to run at least a couple of hours before your analytics tables run. That way you will be able to see the predictor data elements in your analytics. You could also put this in validation rules to produce outbreak alerts.
Hope this helps.