Next, I want to show you an example of a data set that has no numeric fields whatsoever but we could still summarize it with a pivot table. We need to close this file. So just choose File, Close and then I will open the file for 02 pivot tables. This data set has defect data. It shows the date, the manufacturing line number and the area where the defect was found and we want to do an analysis to see which areas have the most defects. Unfortunately, there is no field for us to sum. I will select one cell in the data, use insert pivot table, and click okay. I want to do an analysis by which defects occurred most often. If I choose the defect field, it is automatically added as a raw label.

I now need to take the defect field and drag a second occurrence into the sum values. Because this field is text and there is no way to sum, Excel will automatically choose to count which is perfect and that is exactly what I want. Now, as I look down through the data, I see that fit and finished roof had the most defects of all. So, I want to be able to zoom in and take a look at fit and finish roof. Of course it makes sense to take the defect field from raw labels and move it up to report filter where we can choose fit and finished roof.

Let us do an analysis by date to see if this problem has been happening all month. And now, instead of percentages, I will see cumulative year to date figures. You will see that for most of the month, there were just a few occurrences each day. If I scroll down though, something happened around the 28th of the month. Let us take a look at the line. I am going to take the line field and add that to the column labels. The problem seems to have happened first on line two in column C. That is the line that started having problems. Using this pivot table, it was very easy for us to zoom in and realize that the fit and finished roof first happened on line two and then spread to the other lines a couple of days later.

I bet if we check the records, we will see that a new shipment that arrived from the vendor was restocked to line two and then eventually out to the other lines. This is a very interesting situation where we had many rows of defect data. Really no idea on what is going on but by using a pivot table in the count option instead of summing, we have been able to zoom in and find the problem very, very quickly.