How to sample data to classify sales performance data?

Question

I have data about sales in a year, call and appointment records, and background of salesmen. I want to apply machine learning and data mining to predict which kind of person would bring highest sales each month. My problem is I am not sure how to sample the data. Should I treat sales of a salesman in different months as individual records? or should I only pick one record for each salesman?

Answer 1

If the month-to-month variation is known not to be significant then, for each salesman, compute the average monthly sales and let that be the record for the salesman. If there is some kind of seasonal effect this would not be a good approach as you would really prefer to have January's average sales February's etc. In fact, you can look for seasonality by computing for each month the average across salespersons.