Resolve Costly Out-of-Control Processes More Confidently with Control Chart Triage in JMP Live 19

Hello. Thank you for joining us for this JMP Live Talk. My name is Aurora Tiffany-Davis. I'm a software developer on the JMP Live team, but I'm not alone here today, and that's because JMP Live is a collaboration tool, and the feature we'll be showing you today is a collaboration feature. That works a lot better when we have two presenters, and I have an excellent co-presenter today in Annie Dudley.

Thanks, Aurora. Hello everyone, I'm Annie, and I am a software developer in the statistics team, and I'm responsible for the control charts in JMP. Today, we'll be talking with you about control chart warnings in JMP Live and the extra benefits that you get when you put your control charts into JMP Live. We have been working on this feature over many versions of JMP Live. We've been investing in this effort because we know that if you're using control charts, you're likely using them to study the stability in your process. When your process is not stable, you're relying on control charts to make important decisions, and failure to act on these decisions can be stressful and costly.

Our goal is to help you share information, communicate and collaborate across your organization, and make these decisions more quickly and efficiently. Now, I will create a control chart in JMP. Here we have some data from a ceramics organization, and they are specifically making silicon carbide ceramics. Today, we're going to look at the flexural strength, and we also have the dates for each observation.

I'll create a simple control chart using Control Chart Builder. All of these dialogs launch Control Chart Builder. I'm going to create an individual moving range chart with the flexural strength as the Y and the date as the subgroup. Now you may notice that it has a little note saying, the limits were imported from a column property. That's because I am in what's called Phase 2 of my control chart.

I've been running this process for a while, and so I know what limits to expect. As a little tip, this is going to help you with Version 19 to triage your warnings and JMP Live. Those limits, as you can see, were imported from a column property. I don't need to see that, the limits table anymore, so I'm going to turn that off. I am going to turn on some tests. The first test I'm going to turn on is One point beyond Zone A, also known as one point beyond the 3 Sigma limits. That's everyone's favorite test. I'm also going to turn on show you how to turn on the tests in the control panel.

If I scroll down here, you can see Warnings 1 and Warnings 2. Warnings 1 is associated with the top chart. In this case, it's the individual's chart. Warnings 2 is associated with the moving range chart. I want to also turn on some trends tests, because those are helpful at picking up a trend before it goes beyond the limits. I'm going to choose Test 2, which as you can see is nine points in a row on a single side of the center line. I'm going to turn on Test 5, which is 2/3 points in a row in the outermost zone, and the one that's just inside the limits.

Fortunately we didn't see any more tests flagged other than that first one that we saw. I'm also going to turn on the only test that's available for the moving range chart. I think I'm pretty good. I'm pretty pleased with this now. We do see these tests here signaling, but I am ready to upload these charts to JMP Live now. To do that, I go under File and choose Publish. I've already set up my connection, so I don't have to do that. I'm going to publish my reports to JMP Live. It's grabbed the name of my data table in my output window. I'm going to click Next.

Now, I need to choose a space to publish this to. Since we're at discovery, I'm going to publish it in the Discovery space and specifically in the Europe 2025 discovery space. I can click Next. That looks pretty good. I'm not going to add a description this time. I go ahead, and I'll click Publish. Now I'm going to turn it over to Aurora, and let's see what she sees on her end.

Thank you, Annie. Yes. Here I am on JMP Live, and we are going to take a look… Oh, see new posts. Here's the one that Annie just published a moment ago. I promise, we are going to look at this report that Annie just published. But just in case there's anybody here looking at this talk who hasn't seen JMP Live at all, I want to take a moment and just give you a very brief introduction to the JMP Live product.

Here we are on JMP Live version 18, and I'm going to show you this one first, because that's the version that is available today. JMP Live is a secure collaboration platform. It's installed on your own servers, and JMP Live is used to share JMP content with your colleagues, whether or not they have JMP themselves.

Here on the homepage, I can see all of the reports that I have access to, including the one that Annie just published. When Annie published this report to JMP Live, you may recall that she said, "Okay, well, what space do I want to publish this report to?" JMP Live is organized into collaboration spaces, and you can see those if you click the Spaces button here on the top. Here, I can see all of the spaces that I have access to.

Now, you and your colleagues decide which collaboration spaces to set up and who has access to them. I can browse around inside collaboration spaces like so, and when I find a report that looks interesting to me, I can open it up. When I do that, I'm not limited to just looking at the report, but in most cases, I can also interact with it, whether or not I have JMP on my machine. In this example, this report has a column switcher, so I can interact with it by switching from column to column and seeing how that graph responds.

That was your 30-second introduction to JMP Live, and now I'd like to talk about some of the extra benefits that you get from control charts when you publish them to JMP Live. These extra benefits are available for any report that you publish that is based on the Control Chart Builder. But as you recall, Annie didn't click this button when she was setting up her report. Instead, she used the IMR Control Chart. But as she said, that's also based on the Control Chart Builder, so is XBar, so is Short Run, all the way down to and including the Rare Event Control Chart. Everything from here up is based on the Control Chart Builder, and so all of those will get the extra benefits in JMP Live.

What are those extra benefits? One of them is that we help you find reports that have warnings in them. Wherever you are in JMP Live, if a report has warnings in it, it will be accompanied by this red icon saying there are control chart warnings in this post. Also on the homepage, you can filter if you want to see just those reports that have warnings in them. When I click this checkbox here, I get a filtered list of reports, and it's just the ones that have warnings in it.

When I open up a particular collaboration space, I have over here on the left-hand side this Warnings panel, and this shows me just the reports that have warnings in this one space. You can also get notifications about control chart warnings. Here, I have a notification about the silicon carbide ceramics report that Annie just published, and I also have an email in my inbox about this as well.

These notifications and the emails, they don't get blasted out to everybody in your company. You get a notification like this if you are the publisher of the report, if you're in a group that an admin has set up to get these notifications, or if you yourself decide to opt in. You can subscribe to control chart notifications about one particular report, about a folder on JMP Live, or about an entire collaboration space on JMP Live.

Let's follow this notification link to take a look at the report that Annie just published. Just like in JMP, you can see the warnings highlighted in red here in the report. Just like in JMP, you can hover over a particular warning to learn a bit more. In JMP Live, you have this Warnings panel on the left that summarizes what's going on with the warnings in this report.

You've got a warning rate of 2.5%, and you can see here that Annie has chosen to apply tests 1, 2, and 5 to the individual report… I'm sorry, the individual chart, and Annie has chosen to apply tests beyond limits to the moving range chart.

For the remainder of this talk, I'm going to focus on the additional benefits that you get above and beyond JMP Live 18, the ones that we've added into JMP Live 19. To summarize those additional benefits in 19, we want to help you find out of control processes more easily. We want you to learn more about the warnings in a process. We want to help you work through those warnings one by one, and keep track of your progress while you're doing that. We want to help you stay focused on just those processes that still urgently need your attention.

Disclaimer here, when I switch over to show you JMP Live 19, keep in mind that JMP Live 19, of course, is still in development. To show you some of the changes between 18 and 19, I'm going to show you a few things side by side.

Here is the report that Annie published in JMP Live version 18, and here is that same report in JMP Live version 19. One difference that you notice right away is that the report is bigger. That's a minor thing. We're just using all of the space that we have available to us, and that's not specific to control charts. But otherwise, 18, 19, they look pretty much the same. You still have the warnings highlighted in red. You can still hover over a warning to learn a bit more. You still have a Warnings panel over here on the left.

Let's say that I'm a quality manager, and I'm expected to really know what is going on with these warnings. What can I learn here? In version 19, I can still learn the stuff that I learned in version 18. I can see a summary for each chart. For individual, I can see that the alarm rate is 2.5%, and that Annie has chosen to apply tests 1, 2, and 5. Whereas on the moving range chart, Annie chose to apply tests beyond limits. But I have some new things here as well in the Warnings panel in version 19.

We have two warnings in the report, and so we have two warnings listed out individually in the Warnings panel. You can see more information for each individual warning. You can see which subgroup failed. You can see the actual measured flexural strength at the failure, 389 MPa. You can see that it was specifically test 1 that failed, even though tests 1, 2, and 5 were all applied, this is the test that failed.

You can work your way through this list of warnings in two ways. First, you could click through this list in the Warnings panel on the left, and when I do this, keep an eye on the warnings in the report. You can see when I click through the list in the panel, we change which one is highlighted in the report, and that's just to help you stay oriented between the list and the report, so you know exactly which one you're looking at. You can also work the other way by clicking through the warnings in the report, and there as well, we keep you oriented by showing you which one you're looking at in the list.

I can learn a lot about the warnings in a report, but what can I actually do about them? As a quality manager, ultimately, I'm responsible for making sure that each of these warnings is dealt with, and I need to work with specific people to make sure that these things get investigated, and hopefully fixed and even documented.

Starting in JMP Live version 19, we're adding this idea of triaging your control chart warnings. Let's try that out. Here it is in the Triage section. Let's take a look first at the warning that we got in the individual chart, where the flexural strength of 389 MPa was below our control limits. That's pretty concerning, and I want to make sure that that gets investigated.

What do I have here in the triage area? First, I can change the status. What is the status of this particular warning? It could be open, investigating, or addressed. I really don't know what's going on with this one yet, so I'm just going to leave that in the open status. Second, assignee. Who is responsible for this warning? Somebody needs to be responsible for it. I could assign it to myself with one click. But honestly, I think that Annie is going to do a better job of investigating this than I am, so I'm going to assign this to Annie.

Next, we have notes. In other words, what is going on with this warning? I really don't know yet, but I do want to be polite, so I'll leave a note here for Annie. "This needs to be investigated. Thank you." By the way, when I assign this to Annie, I'm not doing that because I know that she has JMP on her machine. That really doesn't matter. I could assign this to anybody in my company who has access to this report on JMP Live, and they could take part in this investigation process, even if they don't even know what JMP is.

I'm going to assign this one to Annie, and while I'm here, I think I'll also assign her the warning from the moving range chart as well. I'll leave a note too. "This needs to be investigated. Thank you." Now let's see what that looks like for Annie when something is assigned to her.

Thanks, Aurora. I can see that I just received some notifications, and I just got an email notification saying that Aurora just assigned some warnings to me. I can see it here in the notifications list that Aurora just assigned the warning, and when I click on that, it takes me directly to the warning itself. But what if I didn't pay attention to that nice notification bell? What if I've been sending all my emails to the to spam folder? Can I find them in some other way?

Let's start by going home to the home page, and I can start by looking at Control Chart Warnings. This filters down on all the charts, all the output up here that has warnings on it, and then I can choose warnings assigned to me. Give me some stuff to work through. I see there's only one chart that has warnings assigned to me. It's kind of a relief. From within this chart, I can see there are two warnings, for different subgroups, that have been assigned to me. Now, let's work through these warnings.

This first one, you notice when I click on here, it highlights it over here. I happen to know what happened when this particular strength went down. I'm going to type in the notes here that, "This was a silicone sand mixing issue, and it's been fixed." I'm going to change the status to addressed, because we did. We caught it right away, and as you can see in the chart, it's been fixed. Click Save there, and let's open the second one. Now you notice this one, this is the moving range, that means that's the difference between this point and that point. That makes sense that it was that it was also high. If we look at this point-

Annie, click the little caret.

Thank you. This is the same problem. I'll say, "This was also the silicone sand mixing issue, and it's been fixed." I'll click Save here. Because I clicked Save on both of these, you notice that instead of seeing a red circle now, you see a green square, and you see the little green circle here on both of these subgroups, both these warnings that were showing up. You also see the green check mark next to the Warnings tab. I knew what was happening.

Now I have an effective way of communicating with not just Aurora, but everyone, all the colleagues across my organization on each specific warning, specifically, whether it's dealt with and what its current status is, who it's assigned to. I was responsible for this one, and so I'm going to leave it assigned to myself. Any notes on what was going on. Not only am I able to see it myself, I can share it with Aurora, but anyone across my organization can see this, and they can see it real time.

Now I'm going to turn it back over to Aurora to go over more of the bells and whistles associated with this.

Yes. Thank you, Annie. As Annie said, this all updates real time, so as she was addressing those warnings, I was seeing those updates happen here on my browser. I can see that she's dealt with them, and I can see that in a few ways.

As she said, you can see it because they changed from red to green in the report here. But I'm the kind of person who always likes to double-check things. I'm going to hover over one of these warnings, and I can see here in the hover help explicitly, this is addressed. It was dealt with by Annie, and it was a silicon sand mixing issue that has been fixed. That's great to see.

I can also see that these warnings are dealt with because I have these happy green check marks here in the warnings list. But again, I always like to double-check things. I'm going to filter the list of warnings that I'm looking at for this report. I'm going to filter them to exclude the addressed warnings. When I exclude the addressed warnings, there's nothing left. That gives me a really warm, fuzzy feeling that this process has been dealt with, and it should be under control again.

But I'm really interested in the fact that we had flexural strength that was measured below our limits. That's really quite concerning. I want to know exactly what happened here. I want to dig into the history of what happened with this one warning. I'm going to open that particular warning and click on Open History. Here I can see the full history. I can see who assigned it, and when, and what they said. I can see who dealt with it, and when, exactly when, and what they said. Now, this was a really straightforward fix. It got assigned to one person who immediately addressed it and left helpful notes. But sometimes it's not that simple, and there's a bit of back and forth. If there had been a more complex history here, that would all show up on this dialog.

I can also download this history of a particular warning. I'm going to open that file that I just downloaded in JMP, because I like JMP. But it's just a simple CSV file, so really, you could open it pretty much anywhere. In this file that I downloaded that describes the history of this one particular warning, first, we tell you an awful lot about that warning, and that's to make sure that you know exactly which one we're talking about.

This is the one from the individual chart for silicon carbide ceramics measuring flexural strength. It's the one that took place on January 29, the one that had a value of 389 MPa. We want to make sure you know exactly which one we're talking about. Then in the bottom of the file, we get into the history. You can see when this warning came into being, when I assigned it to Annie and left her a note, and when Annie addressed the problem and updated that note.

Speaking of downloads, I can also download not only the history of one warning, but I can download the current state of all of the warnings in this report. When I do that, I could choose to filter it. I could decide only to download warnings that are assigned to me, or I could choose not to download the addressed warnings, but I don't really want any of these filters, so I'll just download the basics.

Again, I'm going to open this in JMP because I like JMP, but it's just a CSV file, open it wherever you like to. Now, this file looks a little bit different. It has one row for each of the warnings in the report, and you know there's two warnings in this report, and that's why we have two rows. You can see that both of the warnings are addressed. They were both dealt with by Annie, and they both have some helpful notes. Now, the rest of the columns in here are describing each of the warnings in detail. There's the one from the individual chart and the one from the moving range chart. We've got a lot of columns here, because we know a lot of stuff about the warnings.

But really, this download, to be honest, it looks a little bit repetitive, because both of our warnings were from the same JMP data table. They were both from the same column in that data table. Ours is not a chart that has phases, so we have a pretty simple example, and that makes this download look a little bit repetitive. But you can see here that we're set up for much more complex scenarios, where you have multiple tables, multiple columns, multiple phases, et cetera.

I've dug into the history, and I've got an overview of all the warnings in the report. They're all addressed. I've got green check marks everywhere. I'm feeling really good about this process, the silicon carbide process. But what other processes should I be worried about as the quality manager? One of the extra benefits that I showed you in JMP Live version 18 was the ability to find reports that have control chart warnings. We also have that ability in version 19, and we think that we've improved it. Again, I'm going to show you 18 and 19 side by side to refresh your memory.

Here is the home page in version 18 of JMP Live, and here's the home page of version 19 of JMP Live. One of the changes you can see is that we've added these sections here so that you can find certain things at a glance. One of these sections is Control Chart Warnings.

Now, this section shows you a handful of the most recently updated reports that have warnings. If you want to see all of the reports that have warnings, click on View All, and here you can see all the reports that have warnings. Here's another thing that's different between 18 and 19. Back on version 18, every report that had warnings had a red icon. Now, in version 19, we have red, yellow, and green. What's going on there?

With the ability to triage individual warnings in JMP Live version 19, comes the ability for JMP Live to keep track of your progress. The one that we've been working on has made great progress, and it's showing up with a green icon indicating that it's fully addressed. Annie has moved every single warning in this report to that addressed status. This is a signal that things are pretty much under control again, and people shouldn't be too worried about the silicon carbide ceramics process.

This one down here, though, is under investigation. It's got a yellow icon. That means that all the warnings in there have at least made it to that investigation status. This one here is red. It still urgently needs attention, because there's at least one warning in that report that's still in that open status. Nobody has triaged it.

We also have some filters. In version 18, as you recall, we had a very simple filter that said, "Do you want to filter down to just the reports that have warnings?" In version 19, we have a filter that gives you a bit more control. You could filter down to just those reports that still urgently need attention, in other words, the ones with the red icon. You could filter down to those that are under investigation, in other words, the yellow icon. Or you could filter down to the reports that are fully addressed, in other words, the ones with the green icon. Or you can do any combination of these that you might want to look at. We also have this filter in a lot more places now than we used to. Pretty much anywhere you can see a list of reports, you can filter down to these control chart characteristics.

Another benefit of keeping track of the overall triage progress for each report is that we can stop sending out notifications as soon as a report leaves that urgently needs attention red icon state. Let's say that you have a report on JMP Live that's got warnings, and JMP Live has already sent you out a notification about that. Now let's say that that report is updated. JMP Live at that point is going to take another look at your triage progress. If you're under investigation or fully addressed, JMP Live is going to assume that you're on top of things, and you don't need any more notifications.

On the other hand, if your report is in that needs attention red icon state, you are going to get another notification. It looks like the silicon carbide ceramics report was just updated, just now, and it looks like it has gone from green back to red. That means it's gone back to that needs attention state. Because of that, I have a fresh notification about the warnings in that report. Now, I know that we already dealt with the warnings in this report, so I'm guessing that it has new warnings. Let's follow the notification and find out what happened.

Yes, it does look like we have new warnings that we did not have before. Where did these come from? Our data just got updated 1 minute ago. That new data must have introduced these new warnings. Maybe one of my colleagues manually uploaded a new data table to JMP Live, or maybe this data is set up to use the scheduled data refresh feature in JMP Live that automatically goes out and pulls in new data from a database or a server. Either way, the data got updated.

Of course, whenever data is updated on JMP Live, all the reports that use it get updated automatically as a result, and that's what happened here, and that's why our report is now showing these new warnings. What are these new warnings?

This is a failure of Test 2, nine points in a row on a single side of the center line, and that is a failure of the same test. These are trends tests, and so we're not exceeding our limits yet, but we might exceed our limits soon if we don't pay attention to these warning signs. I definitely want this investigated. It looks like both of these warnings showed up with the same data update. I think that they might be related to each other, and I would like to assign both of these warnings to the same person.

I don't want to have to fiddle around with assigning one and then assigning the other, so I'm just going to draw a box around both of these two new warnings, right click, and I'll say Triage. Now I can triage two warnings at the same time. I need to pick an assignee, and I'm tempted to assign this to Annie, because she did such a good job with the first ones that I assigned to her, but I don't know if she's already gone home for the day, and I'm honestly pretty worried about these warnings. I thought that this process was really under control, and now we're seeing this troubling trend, so I want these looked at right away.

Annie would do a great job, but honestly, so would any of the other ceramics process engineers. I'm going to assign these two warnings to the ceramics process engineers group, which has five members. Now everybody in that group should get a notification that they are responsible for these warnings. Let's see how this looks for one of those engineers. I'll turn it back over to Annie.

Thanks, Aurora. I can see that I got another notification. When I take a look at the notification, I can see the old warning that I addressed already, and I see these two new ones that were assigned to my group. When I look at them on the screen, I can say, "Oh, yeah. I think we need to investigate that." I don't know what's going on yet. Just like Aurora did, I can draw a box around those and highlight them and say, "Okay, we're going to start investigating those." I can choose Triage, and change the status to Investigating. My only clue is that, "We change suppliers on March 1st." I'm going to also add that, "We're still investigating." Now I've changed those two to yellow, and because both of these are investigating, and that's the worst situation on this chart, the overall warnings notification turns yellow.

We've shown you some things in JMP Live that we think will help you find out of control processes more easily, that'll help you learn more about the warnings in your process, work through the warnings in your processes one by one, or in a group, individually or with a team, with colleagues. For each warning, you can communicate what its status is, whether it's addressed, whether we're investigating it, or if it's still open, and also what you think might be going on, so you can keep track of your current progress on each process. This should help you stay focused on your processes that urgently need attention.

What do you need to do with your control charts when using JMP Live 19? These new abilities are there if you want them. You don't have to use them. But if you want the basic benefits, publish reports that are based on Control Chart Builder. If you want to be able to triage your warnings in a control chart, you need to publish the control charts using static limits. In other words, Phase 2 control charts, what I talked about at the beginning. You can set those limits by using a column property like I did when I first published this. You can set the control limits using a limits table, or you can set the control limits using JSL.

In this list of benefits that we've been building up over time for control chart warnings in JMP Live, there's this one little line here at the end called triage warnings. We think this one little line actually represents a pretty big step forward in helping you understand what's going on with your process and helping you collaborate and communicate more effectively across your organization with your colleagues to solve issues.

We did not do this on our own. We would like to thank the other developers on the team, the JMP product management team, and also the JMP software testers. We would also like to thank all of the JMP employees and customers who've tried this out and helped us along the way with feedback. It's all been really helpful. We'd like to invite you to please stay for question and answers and talk to us throughout the talk or throughout the conference. Thank you very much for your time, and we'll open the floor for questions now.