We sell according to 8 or 9 product codes. I wanted to produce a graph that shows the accumulated sales by each product code, by month for 2011-2.
Now, we NEVER have negative sales for any product, at any time - no returns, chargebacks, etc. Therefore, I would expect every single cumulative chart to be convex. And indeed, if I select any SINGLE product code, the resulting chart is convex - i.e. always increasing. However, if I select all the codes, or most sub-sets, the chart resembles a mountain range, with many peaks and valleys. This makes no sense to me. Now, not all products have been offered for the full two years, and some products have zero sales for a given month, but I still don't see how that affects the total
When I change the type of chart from area to a straight table, and export the resulting data, it shows that for every single product code, the cumulative total of each month is more than the month before (that's the simple check in the "POS" category.
If I take the product code, and make it an expression, with just year-month as the dimension, the resulting chart is convex, but it doesn't show the breakdown by product code (i.e. the chart is one solid colour).
I didn't do anything fancy. Dimensions are 1) Year-Month, and 2) Result code, while the only expression was sum(sales) with full accumulation selected . The exported data are attached as an Excel sheet.
Actually, I went back to look at the individual months more closely. for some products, the sales weren't '0', they were NULL. could this be the problem? I'm going to adjust my script to find out.
