Home / Statistical Tools / Control Charts / T Chart / How-To with GroupBy
How-To with GroupBy¶
This walkthrough charts stoppages on two production lines in one run. GroupBy divides the rows by the value of a grouping column and charts each group separately, so each line gets its own worksheet, its own fitted distribution and its own limits.
On this chart the fit is the model, so a line with a genuinely different failure rate needs its own. Both lines below are in control on their own charts. Pooled into one chart, an interval that is unremarkable for one of them is marked out of control.
This walkthrough uses Fit exponential distribution only, as the T Chart How-To does, so that every number on the page is a plain multiple of a group's own average. Start there if you have not built a plain T chart yet.
The data¶
Twelve stoppages from each of two lines, each value the days that line ran since its own previous stoppage.
| Line | Days Between |
|---|---|
| Line 1 | 6.5 |
| Line 1 | 18.9 |
| Line 1 | 11.1 |
| Line 1 | 22.2 |
| Line 1 | 23.6 |
| Line 1 | 1.6 |
| Line 1 | 0.3 |
| Line 1 | 43.6 |
| Line 1 | 7.2 |
| Line 1 | 6.4 |
| Line 1 | 130.5 |
| Line 1 | 15.2 |
| Line 2 | 10.9 |
| Line 2 | 3.9 |
| Line 2 | 6.1 |
| Line 2 | 1.0 |
| Line 2 | 6.0 |
| Line 2 | 12.2 |
| Line 2 | 4.4 |
| Line 2 | 8.1 |
| Line 2 | 6.7 |
| Line 2 | 0.4 |
| Line 2 | 8.5 |
| Line 2 | 5.4 |
Line is the grouping column. It is not a measurement and you do not check it in Data Columns (required).
Steps¶
-
Put the data in Excel
Press Copy for Excel above the table. In Excel, open a blank worksheet, click cell A1, and press Ctrl+V. Headers land in row 1 and the twenty-four intervals in rows 2 through 25.
-
Start the chart
From the Excel ribbon: QXL Stat Tools Tab > Control Charts > Rare Event Charts > T Chart (Time Between Events).
-
Give it the cells
In the Data Selection window set Selected Range: to A1:B25. Leave Data in Columns selected and First Row/Column is Header checked. Press Next >.
-
Check the measurement column only
On the Data tab, check Days Between in Data Columns (required). Leave Line unchecked: a column can only do one job at a time, and Line is about to become the grouping column.
-
Switch the calculation type to GroupBy
Change the calculation type from Excel to GroupBy. The GroupBy region appears below it.
-
Move Line into the GroupBy order
Select Line in Available Items and press
>. It moves into GroupBy Order as1.Do not switch back to Excel after this
Switching the calculation type back to Excel clears the GroupBy list without warning, and returning to GroupBy means building it again.
-
Tick the exponential box, then Finish
On the Limits & Splits tab tick Fit exponential distribution only, then press Finish. The setting applies to every group.
What you should see¶
Two worksheets, not one, named TChart - Line 1 and TChart - Line 2, each with one panel of twelve points.
| Line 1 | Line 2 | |
|---|---|---|
| Weibull α (scale), which is that line's average interval | 23.93 | 6.13 |
| Weibull β (shape) | 1 | 1 |
| CenterLine | 16.58 | 4.25 |
| UCL | 158.09 | 40.53 |
| LCL | 0.03 | 0.01 |
Both lines are in control. Not one point is marked on either chart.
Line 1 runs about four times as long between stoppages as Line 2, 23.93 days against 6.13, and because every line on an exponential T chart is a multiple of that average, the ratio carries straight through: Line 1's center line and upper limit are also about 3.9 times Line 2's. Each was computed only from its own twelve intervals.
Line 1's longest run is ordinary for Line 1. It is 130.5 days, against Line 1's own upper limit of 158.09, inside with room to spare. On a line that averages 24 days between stoppages, four months of running is a good stretch and not a signal.
Things to try next¶
-
Run it without GroupBy. Set the calculation type back to Excel and finish again. One worksheet, twenty-four points, one distribution for both lines:
Value Weibull α (scale) 15.03 CenterLine 10.42 UCL 99.31 The pooled scale lands between the two lines, as it must, and now Line 1's 130.5-day run is marked out of control, 31.2 days above the pooled upper limit of 99.31. Nothing about either line changed between the two runs of the analysis. It is a signal against a rate that is the average of two lines and the true rate of neither, and it is the only marked point on the pooled chart. Line 2, meanwhile, is being judged against limits nearly two and a half times wider than its own, so a real deterioration there would have to be severe before the pooled chart said anything.
-
Compare the two scales directly. They are on each worksheet's summary table as Weibull α (scale), and with the shape at 1 the scale is just that group's mean interval. One number per group is the whole model.
- Clear the exponential box and run it again grouped. Each group is then fitted separately for both parameters, so the two sheets can come back with different shapes as well as different scales. That is a different comparison from this one: it asks whether the two lines fail in the same pattern, not just at the same rate.
- Compare with a split instead. Split Control Limits also gives separate stretches their own fit, but it divides one chart by position rather than by the value of a column. Use a split when one line changed over time, GroupBy when the rows describe separate lines.
- Group by two columns. Add a second grouping column and it becomes
2.in GroupBy Order, subdividing what1.produced. Each combination needs at least three intervals of its own to be fitted at all.