I have created a database which logs issues as they are submitted by our users. These issues are related to specific job items which they have been assigned to complete, and n this sense each issue has an identifier and each job item has an identifier.
As far as the job items are concerned, the time user is alerted to the item assignment is logged.
Likewise the time of the reported issue is logged, and also the corresponding job item identifier is logged along with this record.
What I am aiming to create is an equation which will calculate the chance of issues occuring considering the current frequency and count of assigned items.
For example (made up example):
12 items assigned on Tuesday, and for this day 2 issues were reported
72 items were assigned on Thursday and there were 32 errors
In this case I can see that the relationship between the number of assigned items and issues reported is not linear, and I am finding it difficult to understand how I can predict the number of issues expected for a certain frequency if items assigned.
Considering I have easy access to historical information linking the number of assigned items with the number issues reported over time is there some method I can use as a foundation for doing my calculation?
As a basis, I will have a table (View/Stored Procedure generated from SQL Server 2005) as described in the code snipped area Table A. There will be multiple records in this.
Could anyone recommend how I could harvest this information to retrieve variables to progress towards what I need, which is essentially the answer to the question:
"If X items are being processed today then how many issues can I expect?"
I'm not sure any answers need to be in programming language immediately as I have not finalised what my development language would be, but if there are any SQL Server 2005 queries and generic formulae offered I would be greatly pleased to see them!
DATE(DAY) ITEMS FREQUENCY ISSUE FREQUENCY