15 Things You Didn't Know You Can Do in Control Chart Builder (2025-US-30MP-2465)

I'm Annie Dudley. I'm a developer responsible for control charts in JMP. Today, I'm here to talk with you about 15 things that you may not realize that you can do in Control Chart Builder. Let's dive in. First, I want to talk about some of the changes in version 19 regarding the legacy charts. They have now been removed from all the menus, the scripting index, and the start menu.

The JSL will continue to run for 19 and probably at least 20. We're trying to encourage everyone to switch over to the Control Chart Builder. But if you really want to pull them back, there is a preference to bring them back onto the menus, and it's under preferences under the menu option on preferences, not under the control chart option.

Let's talk about Control Chart Builder. As the legacy charts are being retired, I've been working on adding more and more of the features from the legacy charts into Control Chart Builder. Let's talk about some of those features.

All right. First, let's talk about the dialogs because Control Chart Builder has a bunch of dialogs now that can create the same types of charts that you could with the dialogs with legacy control charts.

If I go under Analyze, several releases ago, this menu brought you to the legacy charts. You notice there's no legacy option on this menu anymore. But we have all these menu entries into Control Chart Builder. If I take a look at, say, the standard individual moving range chart, we can now see a K-Sigma, a range span, options to change which dispersion chart you want for your charts. There's a get limits and capability option on here.

Let's take a look at, say, a summarized chart, like an Xbar and R chart. Again, if I enter, say, diameter, you notice you can specify a constant subgroup size on the dialog if there is no subgroup column given. I do have a subgroup column. I'm going to enter the day as my subgroup. I can also select range or standard deviation if I want. These options that you see on the dialog are going to default to whatever your preferences are. The preference is currently set to range.

Again, you can pull in a limits file right here. Let's go ahead and launch this. You notice it's launched, Control Chart Builder, and it's using the variables that I specified on the dialog. There is one more dialog that I'd like to talk with you about. Well, actually, two more. Where there's the run chart, some people prefer to see the center line on the run chart when it's first launched. Some don't. That option is there.

Also, because each and every one of these items, by the way, I am talking about because people, customers, you all, have asked me or some of the other JMP folks if these options are available. I wanted to go through each and every one of these.

As we look at the different dialogs available, you see there is a dialog entry for Levey Jennings Control Chart, and that is going to create a Levey Jennings Control Chart.

Now, let's go back to the setting the case Sigma. I showed you how you could do it on the dialog. If you skip that part, you can also set it under the red triangle menu. There's an option here to set K-Sigma. Say, for example, I want to change that to four and have wider limits. You notice that the limits are now wider, and the control limits are reflecting that right here.

I can also set Sigma. There's an option here to set Sigma. It defaults to the Sigma for this particular graph. Maybe I want it to be a little bit larger, say 0.3 instead. This is going to additionally widen our limits. It gives us a little note here that the limits were computed using a specified Sigma.

Now, the next thing I wanted to talk about was setting the limit labels, showing them on the graph. In version 19, I've organized, I think, all the limits commands under this Limits submenu so that you can gravitate toward that if you want to get, save, or show the limit summaries and so forth.

Let's look at showing the limit labels. When I select that option, the limit labels, the graph scoots over just a little bit, and now we have the limits labels on the graph. Some people have asked, "Can I change the precision of that?" Just like you could in the legacy control chart platform, there is an option to change the precision. Again, it's under the Limits menu, and we can say, "Look at Limits Label Precision."

Now, the way that the limits precision works is the same way it worked in the legacy control chart platform. That is, it takes the precision that it sees on the Y-axis, and then it adds whatever number is in this box to give you the precision. There's one decimal point on the Y-axis right now. We see a two in this box, so we're going to have three decimal places shown. Let's say I want more. Let's say maybe I want five decimal places, I'm going to enter a four here, and that will bring me out to five decimal places.

Box Plots. Some people have asked me, "Hey, can I show box plots on my chart?" Yes. There's two different ways to do it. Box plots are based on the data. Anything that's based on the data is going to be under the Points menu. That can be found here when I right-click, or it can be found if I open up the control panel, then we can see the box plots again appear under the Points menu. It's not going to appear under the second points menu because it's only going to appear for summarized location charts. While it's summarized, it's not a location chart.

All right, so let's move on to short run charts. Now, short run charts were new in version 18 in the last release, but they didn't get a whole lot of PR. Let's talk about short run charts. This concept is where you're making… The whole premise behind short run control charts is when you're making more than one product on a very efficient single production line. This is data that I simulated, and we're making, in this case, chocolate. It's a very fancy chocolate factory in that we make each box of 12 bars whenever we get an order. It is really, really super high-end chocolate. We can see here I have a box of 12 extra dark, and then we have a box of dark, and then followed by a box of milk chocolate and so forth.

We want to monitor our process. To monitor that, I'm going to use the Short Run Control Chart menu entry. I want to monitor my cocoa solids because I'm making extra dark, dark and milk chocolate on this production line. I want to make sure that if somebody's ordering dark chocolate, it doesn't taste like milk chocolate. If it's dark, you don't want it to be way high in the cocoa solids or really low because then your customers aren't going to be willing to pay all this extra money for this very fancy chocolate.

I have my boxes, my subgroup, and then you need to specify which product part so that we have an identifier for it. It defaults to the scaling method of centered. You can also switch to standardize. But to start with, the default is centered. I'm going to click Okay. We're switching types of chocolate pretty fast. We're switching our products pretty fast. We have very short runs of one or two.

By default, I'm showing the product separators on here, and that's a little bit busy. I'm going to turn that off. But our points are colored by which product it is that we're making. We can see here that the milk chocolate is blue, and the dark chocolate is red and that the extra dark chocolate is green.

We can see that we have some issues here going out of control, and especially on the range chart, which is really worrisome. I can right-click and turn on some tests. Let's say test one. I also like some of the runs tests, so maybe test two, but luckily that one's not showing up here. The only test available for the range chart is the test beyond limits. But we can all see we've got a problem going on with the range.

But at the same time, we can also see that most of these appear to be milk chocolate. Now, this makes sense because the milk chocolate has different target statistics. We can actually specify those because we know we have goals, we have targets. We can enter that information and set the target statistics. Our milk chocolate has a target of 37% cocoa solids, and our Sigma is four because it can really vary. I think, if I recall, European standards say that the milk chocolate can go down to 34% cocoa solids, and it can go up to 45.

Now, dark chocolate, we're going to aim for 70. But we want a tighter Sigma on that because dark chocolate has a little bit… It can go down to like 65, maybe up to 75, but we don't want it to go in the extra dark range. I'm going to put a Sigma on here of 1.5. Now the extra dark, we have to have a really tight Sigma because we do not want it going all the way up to 100% cocoa solids. Because these all have long tails.

I'm going to enter my target of 85, and we're just going to give it a Sigma of one. Right now, I'm going to click okay, and we're going to see that our chart has changed. Now what is being subtracted, what's being plotted here is every raw value from the data table subtracting off these target values. It's not taking into account the Sigma yet because this is just centered data. It's not standardized.

I can open the control panel and say, "We know we have different Sigmas here. It makes a lot more sense to standardize this." I'm going to change the statistic to standardized, and I'm going to change the points statistic to range standardized. Now we can see the milk chocolate is not causing a problem, but we do have some noise going on with the dark chocolate.

We may want to… This tells me that we may want to pay a little bit closer attention and maybe tighten up our controls on our dark chocolate so that we can keep it more closer to a target of 70 and not going quite so loose on the production.

There's a short run control charts in a brief description. Now, let's talk about setting limits. If I create an Xbar and R chart, you notice that I could choose Get Limits here. We can choose it here. I'm going to show you a couple of other places.

I can set the control limits by right-clicking and choosing under limits, the limits option, set control limits. If I open the control panel, I can set control limits here and this pulls up a dialog. Why don't I go ahead and set these limits? Okay, so we're going to set from 4 to 4.3 to 4.7. Let's make that 3.5, so they're somewhat balanced. Okay. All right. Here you can see these are identified as User Defined. So those are no longer calculated. This makes sense. When we're in phase one of our control chart, we're still gathering data, and we're trying to figure out, "Okay, what does stable look like?"

Then once we've figured that out, then we can apply our historical limits. They can be in a column property. They can be in a limits file. You can save those limits in a column. That's going to automatically put them in a column property. You can save them to a new table, or you can save them to a new tall table, which lets you go back and forth with process screening.

They can be imported from any of those places. When I save out JSL now, it's going to save out the JSL. Let's set the control limits for only the first chart in this example. There's our setting the control limits. Now, let's talk about an alarm script. Alarm scripts were added to Control Chart Builder in version 15. Just like legacy platforms, in order for them to work, I need to first arm the alarm script. We choose alarm script. What this does is it's just going to send a script to say, "Okay, start listening." There are these particular variables that control charts, both the legacy and Control Chart Builder output, that you can manipulate in your JSL, so like qc_test and qc_col and sample and so forth.

All that's happening here is a big substitute for that. I'm telling it to write with explanation, and now I need to arm it, so I can turn on some warnings. I'm going to turn on Test 1 here, and I'm going to turn on Test Beyond Limits there. Now, if I view the log, you can see it's taken that script that I sent and has given me detailed information about all the warnings who are signaling. I can see that there's one point beyond zone A for sample 11. I could get more information, but I just went with the default script.

There's our alarm script. Now, we're getting into some more fancy things for 19. Okay, so previously, when you created a control chart in version 18 and before, the Control Chart Builder… I'm going to go through the dialog here or the interface. Control Chart Builder would aggregate all of your X values into… Okay, so if it's males, we're going to aggregate everything into males, and if it's females, we're going to aggregate it into females. People have requested that that not happen because they want to be able to show what's going on.

If we have, for example, in this case, we have Blend Time, and I just pull in blend time. Right now, it's colored and shaped by the last column of whether the lot was accepted. I want to see a relationship between the blend time and some of the other variables. I can do an individual moving range chart on this. To see the individual values, I can drag compressor down to the label axis.

Now, the label zone will, and then I can change the increment to one, and then it shows me everything. I can see all of that, but I would like it aggregated in the order that it appears just like the legacy control chart platform did. Then I want to be able to do an individual chart on the means. To do that, let's see, I'm going to set that aside, and we're going to start over. I'm going to… Let's just do it through the interface. Let's say we've got blend time here, and I am going to pull compressor down here. I don't care I want to remove… Also, I should point out there are lots of different ways that you can remove individual charts.

If I right-click on the Y-axis, I can choose Remove. Then I can choose which one I want to remove. I want to remove the R chart. I don't want to look at the Xbar chart here. I want to change this to an individual on the means. I'm going to change that to a moving range. Now I have aggregated the points, it's similar to this, but not quite. I've aggregated the points. Every time, for example, you've got the first compressors showing up. This is those first three rows right here. This is going to give you the same type of chart the legacy chart would have given you. If you'd chosen, I get, pre-summarized or Xbar R. Now you can get that type of chart. I can get what Control to Builder used to do, where it aggregates everything and doesn't pay attention to the row order. That's what version 18 and before would do.

Then I can switch back to the default. Here you have the sort by row order working. Okay. Flip the name, called it Sort by Subgroup, and change it so that this is now the default.

Okay. Let's see. All right, here's an example of that. I also showed you the duplicate values in the X-axis with the label role. There's also an option to use excluded points on the Moving Range, which is brought over from the legacy platform.

Lastly, I would like to show you how this additional new feature in version 19 works, and that is the Show Last N Subgroups. A bunch of you have been asking me for this. Now, this was not in the legacy platform. This is new. Let's just create an individual moving range chart on diameter. We've got a number of observations in this data table, 240, I think.

Now we have our great data table our great chart here, things are looking pretty good. But as I add more data, it's going to get really busy, and it's going to be hard to identify things that, say, for example, are going out of control. We want to be able to really zoom in. I'm also going to turn on some warnings. We want to be able to zoom in on the most recent things that are going on.

Okay. There's an option to Set Last N Subgroups. Now, since right now we have 240, I only want to show the last 50. What this option is going to do is it's going to… Because I'm estimating the limits. I don't have fixed limits right now. We want all the data on the graph. We don't want to subset it, or we don't want to use a local data filter because that's going to change the limits.

If you happen to be in this state, all this is going to do is it's going to shift the axis. Now, let me show you. I'm going to turn off this limits summaries so that I can have the data table next door. All this is going to do, and I'm going to try adding some data here. Let's see. Let's get this in a little bit. There we go. 4.1, you can see it shifting.

I meant to hit 5.2. You can see that the graph is going to continue to shift as new data is added. The points are still going to be tested with whichever Westgard rules or Western electric rules you've selected. This just shifts the graphs to the last however many you select.

This concludes my list of 15 features that you may not realize you can do in JMP. I want to thank you all for coming, and please come talk to me and ask questions about Control Chart Builder whenever you see me at the conference. Thank you very much.