2025-Lean Six Sigma GreenBelt Control Phase-Statistical Process Control-SPC

1. Introduction: I welcome you all to my new class on data analytics using Minitab. In these series, we are going to talk about control charts. As I promised you in my Telegram channel that I will be launching the control chart chapter in detail so that all the basic doubts that we have, we'll get clarified. So the focus of this entire program is to understand what is control charts, what the different types of control charts that we have. Ventura use control charts. What our control limits, what is the central lying in a control chart? Using control charts to detect common cause and special cause variation. Using test for special causes and control charts. What are the different types of tests which one to use when, what to be avoided? Everything will be covered in this session. Who is this class for? This is an important question which many participants have that should I be attending this class or it's not for me? So if you are a student of Lean Six Sigma, Green Belt or Black Belt or a yellow belt. Or you are one who wants to understand statistics, process control, that is about control charts. How can I monitor my process statistically? You want to apply statistics, you want to understand different graphical techniques, grab all you want to understand one of the most important Seven QC tool, which is control charts. You must take up this program because it is good to clarify all your doubts. Some of the key takeaways which you will have from this workshop is you will understand exactly what is control charts. Venture I use it water control limits. What's the difference between specification limit and control limit? And how do I test for special causes? What are the tests that are used for identifying common cause? And what are the different types of mistakes people make when they are working with control charts. All these topics will be covered in detail. I'll be setting up a separate workshop, which we'll be talking about. Practically doing all of it. I'll be uploading my project data sheet. You will get lot of projects to work on. Plus, if you have any queries which are relating to your process where you're drawn control chart but you're not getting an answer. You will free, feel free to write to me or put your comments in the discussion section and I'll help you out, right? So let's get started. 2. Class Project and how to upload it: We need to do a project whenever we are learning something. And the best part but Skillshare is that it encourages teachers to create projects which the participants and the students need to learn from the course, apply the concepts. So I'm going to attach this Excel sheet as a project plan. This has lot of data example that I'm using during the course, like the cam shaft length data, umbrella data, injection molding data's still lend. There are a lot of examples which are available over here, which will help you understand the concept. Number one. Number two, you have this data which can help you do your project and complete it. Now some of you might be new book Skillshare to help you understand how do I do a project and submit a project, I'm going to guide you. So first of all, you would have a datasheet like this. This is the product datasheet which I'm going to share with you. So for example, I am taking my umbrella p data, write it directly, takes me to that place where my data is saved. I copy this data to MiniTab and do my exercise. I have my data. What type of data use case we will be covering it in detail. I'm just trying to show you how do you upload your project. So I have created my control charts and I go ahead and defective umbrellas subgroup size. And I click on Okay, once I have my chart ready, what I need to do is I just copy this graph. I can put it in paint. And I save this file. Now, let's come back to Skillshare. In Skillshare, once you complete the project, you have a section below about review, discussion and project and the sources. So click on Project and Resources. And you can see a green color button over here telling create project. You need to click on this Create project. It takes you to a screen like this, where it's asking you to the photo that you want to present in your project. So I will upload a photograph that I just created. It's uploading. Now I submitted the project title. I have done the p-chart using umbrella dataset. You are free to use all the dataset that I have uploaded all the projects because this will give you confidence and I'm going to share feedback with you. But if you have data relating to your own and you want me to help you out, you can still upload it. I'm not going to ask you for the data, but I'll help you in the interpretation. I have. Read this. For defective data. If I have more content, I can go ahead and add more content. If you want to keep your projects private, you can click on it. But I'd say that we all learned by sharing with each other. And then I come up and I have published. Once you publish it, it might take few minutes or few minutes to few hours for the project to be available. What happens is as a trainer, I get an alert over here that somebody has created a new project and I need to give a feedback. So I go here and give a feedback, which will help me understand what you have done and I can share the feedback. So this will encourage you to ensure that you complete your project. Now let us come back to this chart. Yes. So if you have any questions, you are free to open up a discussion and say start at one position, ask a question on the project. So I have a question. You can ask. Can you help me with more examples of B tilde? I'm just writing it down. So then if you have any questions, I'm just giving a hypothetical question. And I can go down and say, Post. When I pose this, it is available and it is a duty of a teacher to reply back to this. With this, your doubts will get clarified. Whatever you're learning in the class, you will apply. This class is not only going to cover the practicals, but it will also want to clear all the concepts which sometimes gets confusing. Right? So with this, let us continue with our learning exercise. 3. How detailed is this course on Control Chart: Hello friends. I welcome you to this class on control charts, which is a statistical process control methodology. Most of us, when we're doing Six Sigma projects like greenbelt projects or Six Sigma Black Belt project. We are using control charts. Some of the participants are, some of the people in the office also use control charts as a seven Q. C2. It is good that we all know how to practice control chart or how to develop the control chart using different tools. You have templates which are available on a skew. Some people use MiniTab, some people use JMP. In this class. What I'm going to cover is starting from the basics. Though you might be aware of control charts. You might be aware of certain concepts at a high level. I'm going to take you step-by-step in this course, which is going to cover what is control charts? What are control limits? What is a central line in the control chart? How do I detect radiation? Examples of common cause and special cause variation? How can I use brainstorming to investigate into this problem? What type of difference? What are the different types of tests that are available for us? What are the different types of control charts depending upon the type of data that you need to build. We're going to take up many examples where we will try to understand the concepts from the scratch. So we've are going to learn how do I actually apply which control chart will be used? Not only in theory, but also in terms of practical, where I will give you a use case and ask you or guide you through the process of selecting the correct control chart. Post, which I'll also be showing you how to build this control chart using MiniTab and how we identify the special cause variation within the process. Many examples over here, which makes it easy for you to understand and apply it even though you are from different industries. So first I will explain the concept. What is the problem the quality engineer is facing? What is the data he is collecting? And based on the data, which is the right control chart that needs to be picked up. So I hope you enjoy learning through me, not only the practical of creating the control chart, but also about the concepts which are very important for us to ensure that we are able to clear our y-bar during the presentation. The next important thing which I'm going to add is the project. So please watch that in the next video. 4. What is Control Charts: So what is control charts? We have B as we understand, right? It's a chart which helps you find if your process is in control. Audio process is out of control. It helps you identify the presence of special cause variation if it's existing in your process. Whenever there is special cause variation that's present in your process, your process is said to be unstable, which means corrective action is necessary in your process. Control charts are used as the seven QC tool. It is also used to in the control phase of your Green Belt and Black Belt projects. And it's always good to check is my process which I'm running today isn't in control or is it out of control, or are their chances of going out of control? We want to identify the special causes variation, and we want to fix it. Control charts and nothing but graphs that plot the process data in time order sequence. You would have heard about the concept of time series plot. So yes, it is a time series plot or run chart, but the additional element which gets added in control chart, easier central line, upper control limit and lower control limit. And these did, these lines are drawn based on the data that you have provided. The center line represents the process mean. The control limits represents the process variation. By default, the control limits are drawn at a distance of approximately three standard deviation above and below the central line. So when I have a run chart, technically a time ordered chart, but I have additional elements which gives me more information like the central line, my upper control limit, and lower control limit. It helps me to identify ease my process audience, my process is not stable. The points that fall randomly within the control limits indicate that your process is in control and exhibits only common cause variation. So common cause variation could be anything which is a random, right? So sometimes when I'm speaking, they might be a lag of 1 second. That could be because of the common costs. But if I'm speaking and you're not able to hear me, That's a special cause. Either my mic is not working or the my network is not working. I have clicked on the mute button. There will be a special cause which needs investigation. And those can be identified when points fall outside the control limit. Are they display a nod? I'm back then. Even if the process is inside the container limit, but the pattern at non-random, they indicate that the process is out of control. Special cause variation is present in your process, which requires you to take action. When do I use control charts? That's a common question which people have, whether you are monitoring and ongoing process audio trying to obtain understanding of your new process. Control charts is a very helpful tool. You can use control charts to demonstrate whether your process is stable and consistent or what type. A stable process is one that includes only common cause variation and does not have any out-of-control points. Verify that your process is stable before you perform capability analysis. So you remember in the measure phase of your project, you try to do a capability analysis is how capabilities your process if you're doing a Six Sigma project. Some pros, some organizations also do a capability analysis on a regular basis to validate if the process is stable. So you have to do this analysis only after you know that your process is David. If a process is not stable, you should not be doing a capability analysis or that report but you'll get is not a valid report. Assess the effectiveness of the process change. This is also one more scenario where you use control charts. When control charts, it is easy to compare the shifts in the process mean, changes in the process variation. Communicate the performance of your process during a specific period of time. As I told you that can pull charts are drawn in a time order sequence. What our control limits. We will cover this in the next video. 5. What are control limits?: Let us understand what our control limits, control limits are. Your control chart represents your process variation which helps you. I indicate easier process in control or is your process out-of-control controlled remotes, limits and nothing horizontal lines which are drawn above and below the center line. That will help you to judge. Again, is my process stable or not? That person control limits are based on the random variation in the process. By default, Minitab sets the control limits as three standard deviation above and below the central line. This is a simple example, the control limits. So you might have your process data which is being drawn in time order sequence. That red line on the top, which is called as UCL, is your upper control limit. And the NCL is a lower control limit. The green line is your central line. We can see in this process that I have two dots, which are two data points which are beyond the upper control limit. It is indicating that the process is out of control. So it means there is a special cause variation and we need to investigate it. Let us understand the difference between control limit and specification limit in the next chapter. 6. What are central line in control limits?: In this chapter we are going to understand the important terminologies like what is the central line and what are the control limits? So what is a central line in a control chart? The dental line in your control chart represents the process average, not necessarily their desired process. Tonight the specification average given by the customer. It's what is actually happening and happening in your process. The center line is a horizontal reference line on a control chart that is an average value. And it is based on the quality characteristics used the central line to observe how the process performs when compared to the average. If the process is in control, the points will vary randomly around the central line. See this example. You will, the green line, which you see on the screen, is your center line. It's talking about the process average. Each data point could be a subgroup of numbers. So subgroup, it could be a subgroup sample of five elements, ten elements, and so on. And that average is taken and the average of the whole data is taken to derive at this green line, which is your central line. In this example, the X-bar chart displays the length of the manufacturing camshaft over the period of the central line shows the process mean. And the subgroup means very randomly around the process. So you see the dots are going up and down around the green line. And this is happening because of the presence of some common cause variation in your process. Award of caution whenever you're working. Do not get confused between the central line with the target value of your process. The target value comes from your customer specification, right? What is the desired outcome? That is the target line, but the actual outcome is your central line. So both are different. Using control chart to detect radiation. We will cover this in the next chapter. 7. Detect Variations: Let's get started to understand how can I use control charts to detect variation in my process? Control charts are used to monitor two types of process variation. As I told you in the previous chapter, it talks about the common cause variation and special cause radiation. What are common cause variation and special cause variation might be a question that you might have. And what special cause variation looks like when you are working on the control chart. How can I identify special cause variation on my control chart? Using brainstorming to investigate special cause variation is a solution. If you identify special cause variation and your chart, we're going to cover all of it in detail. Do not over correct your process from common cause variation. That's a common mistake which people make. Because some rid of common cause variation should be present in your process because it's natural. So let's understand some degree of variation is naturally in any process. If I start my training, I might be sitting everyday in the morning at ten o'clock to start recording my training. Some days I might be present at tendon. Let's make a more simple example. I order food on swaggy speakeasies that I will deliver the food in 40 minutes. For an example, it might deliver the food in 38 minutes. Next day, I ordered the food from the same place and through 3D, but the order might get delivered in 30 minutes. Third day, it might get delivered in 32 minutes, 45 minutes. Whatever little bit of variation that's there. And it is acceptable that is coming because of the special cause variation. Common cause variation. The common cause variation is a natural process and unexpected variation in the process. Special cause variation is an unexpected variation in the process, which is because of some unusual occurrences. It is important to identify and try to eliminate special cause variation in the process. So for example, I place an order on squeaky. They said it will be delivered in 40 minutes, but it's already two hours and they're not able to deliver my order. I would definitely get concerned. I would not even wait for two hours immediately after 45-minute, I would call up and say where is my order? And they say it's on the way. One hour later means after 60 minutes, I again quite a bit is my order. And they said because there is heavy rains, the driver is stuck on the way and hence, the delay in the delivery. 8. Examples of common-cause and special-cause variation: Hello friends. Let us continue understanding what are the different causes of variation in our data analytics for using control charts, we're using, we're trying to understand all the concepts relating to control cells. One of the important thing that we learn our different causes of variation, which is causing the control chart to move up and down. The causes can be something like a common cause variation or a special cause variation. If you see the Farmer on the right, the boundary farmer on the right, he receives a bag of eggs, which are a box of x, which are a mix of white and colored. When he gets into the root cause of the problem, he realizes that all his eggs, hence our Lee mixed colored x, they are laying white eggs and they are linked colored x, the power. So it's a common cause variation. So he needs to investigate and try to reduce, minimize. You can go, he cannot eliminate material, try to minimize the common cause variation. The powdery farmer on the left also receives a box of x, which are a mix of white and colored. When he investigates into his powder farm, goes to the root cause of the problem. He realized there is a one particular breed of hens which are laying color x rest of the hands and his poultry farm or all Lane white eggs. It means there is an assignable cause to this problem. And hence, he needs to investigate what caused this hand to lay the colored x. It could be at the feed, had a problem, or the bird is from a different place, or it's a mix, there would be some assignable cause to it. And control charts help you identify the common cause variation and special cause variation. Let's take some more examples of common cause and special cause variation. So for example, if a baker is making a loaf of bread, the temperature fluctuation inside the old one slightly baby weigh one centigrade up and down is acceptable and that's due to a common cause variation of the nature of the equipment. But suppose if the temperature is dropped drastically, then it could be due to some assignable cause like the baker forgot to close the window. Special causes help you identify the assignable cause which needs to be focused. One more example is recording the customer contact information. If there is an experienced person, he might be making very minimal errors. Whereas if it's a new person, there is a chance of him making a lot of errors. So it means that he needs to get trained and he needs to be taught some tricks which can help him do this job more efficiently. Let's take one more example to understand common cause and special cause variation in detecting mold into plastic toys. So when the plastic toys and made slight variation, the plastic toys is acceptable because that's the nature of the entire process. You are putting the mold into the, sorry, the melted plastic into the mold and it's coming up. But suppose that the quality of the raw material is bad. The minute the toy is coming out and I hold, it breaks out, it's very hard. It could be due to an assignable cause that raw material is not of the quality that we need and it affects the strength and the consistency of the product. What are special cause variation? And how does it look like when you are working in the chat? A process is stable if it does not have any special cause variation, common cause with always exist, right? Control charts and run charts provide good illustration of process stability or instability is my brother Steven, or it sounds table. We can use, but one of them, the process must be stable before its capabilities assessed or improve our initiated. If my process is not stable, picking up a Six Sigma project is not acceptable, means it doesn't work out. If you'd look at the control chart, which is over here, you will realize that this control chart has some random variation. And these variations are between the two control limits which are identified using the read line. All the dots are randomly fluctuating around the green line, which is my central line. This clearly shows it and it's not even violating any of my eight rules of the control chart. I'll be talking in detail about the different tests that we perform on control charts in the following videos. If your control chart looks something like this with lots of red dots and it says a 11155 and it sees 333. It means that the process is not stable and it has violated the rule number one, rule number five, and number three in this case. And statistics that the process is not stable. There are some special cause variations which needs to be investigated. If you have any questions or any doubts or clarifications, feel free to ask in the discussion section below. Should you try to apply the concepts outside and ensure you complete the project and upload it. In the next video, we are going to learn about using brainstorming to investigate special cause variation. 9. Using Brainstorming To Investigate Special-cause Variations: We continue to understand the different causes of variation in how should I investigate on the special cause variations that we have identified during the, in the control chart. So we can use brainstorming as a very good exercise to investigate the special cause variation. A good starting point in investigating special cause variation is to gather several process experts together, get the subject matter experts over there. Using control charts increases the process operators, the process engineers, and the quality test us to brainstorm why a particular sample. They're out of control. Because you know that when you're drawing the control chart, It's in time order sequence and you will have a sample ID which will identify it. We can go and investigate what happened during that instance. Depending upon your process, you may also include items during the next meeting. When you're investigating special cause variation, you should answer some of these questions. With samples were out-of-control. Which test for special causes did the sample? What does each field test mean? And how do I what do I look for? We're on all the possible reasons for the field test. The common method of brainstorming is to ask questions about why a particular failure occurred. To determine the root cause, you can use the 5-Why method. Keep asking the why, why, why, until you reach the root cause of the problem. You should also use a cause and effect diagram, or the fishbone diagram or the Ishikawa diagram as you call it. To understand the different types of different types of causes which is causing special cause variation. Like whether it's men material method, and so on. Remember, don't over-current your process for common cause variation. But it's important to avoid special cause variation. Try to eliminate common cause variation must make matter worst. Consider a bread baking process, a slight drift and the temperature that is caused by the organs and thermostat up part of the natural common cause variation for the process. If you try to reduce this natural process variation by manually adjusting the temperature, setting up and down, you will probably increase the variable t rather than decreasing it. It is called as over correction. If you have any doubts or need clarification on any of this topic, please feel free to ask in the discussion section below. And as always, try to apply this concept and complete your project. You can dig up some existing control charts that you have and try to find out what is there is a process table. Is there any special cause variation? In the next video, we're going to learn about which test for special causes are included in Minitab. 10. Which tests for special causes are included in Minitab?: Hello friends. Let us understand which test for special causes are included in MiniTab. When we're trying to understand the process stability using control charts. Test number 11, more than three sigma from the central line. This is the most common tests which we always look for. This test identifies the subgroup that have unusual when compared to the other subgroups. These are universally recognized as necessary for detecting out-of-control situations. If a small drift in the process out of interests, you should also do test number two to supplement test number one in order to create a control chart that has a greatest sensitivity. Test number 29 points in a row on the same side of the central line. This test shifts identifies the sift in the process center or variation. If a small shift in the process are of interest, you should use test too, along with test1 to understand what has caused this special cause variation in your process. Because if the process is on one side of the central line continuously for nine bytes, the probability for that point to go out of three Six Sigma. Three Sigma on one side is very natural. And we can understand the reason for it or investigate the reason for it with more confidence. Test number 36 points in a row, all increasing or decreasing. This test detects the trend. This test looks for long series of consecutive points that are consistently increasing or decreasing value. As you can see in this example, if it's continuously increasing for six points, this is also a probability for it to go out of control. And it is already, the process is out of control and going on. It's a special cause. Write something happened in the process which made continuously for the variables or the process to give samples which were consistently increasing or decreasing in value. In this example, it is increasing in value. Desk number 414 points in a row alternating up and down. A test for detect systematic variation. If you want the pattern of variation in the process to be random. But a point that field test for might indicate that the pattern of variation is this number by two out of three points, more than two sigma or two standard deviation from the central line on the same side. So as you can see, there are two examples of the test number five. Over here. This test detects wall shifts in the process very easily. Desk number 64 out of five points, more than one standard deviation from the central line on the same side. So if you see this example, we have four or five ones which were on the same style and they will more than one Sigma away. These desk number six detects small shift in the process like the other tests decently. So number 715 points in a row within one sigma of the central line on either side. Can you see that in this example, these 15 points are very close to the center line. It's as if they are entangling the central line very efficiently. This detects a pattern of variation that is sometimes mistaken as an evidence of good control. This detects control, this test detects control limits that are too wide. Because you would have said this based on your old process control limits that are too wide or often caused by the stratified data, which occurs when the systematic cause of radiation is present in each subgroup. That's number 88 points in a row more than one Sigma from the central line. On either side. This test is called as a mixture pattern. It is in the mixer button. The points tend to fall away from the center line instead of falling near the central line that we saw in the previous test. If you have any queries relating to the different types of tests that you just learn. Feel free to ask in the discussion section below. Ensure that you try to apply the concepts outside and ensure you complete the project and upload it. Thank you. 11. Which tests should I use to detect specific patterns of special-cause variation?: Let us understand which tests should I use to detect a specific pattern in the special cause variation. We learned about different types of tests in the previous video, we should learn how, how can I use them most appropriately? Apply certain tests based on your knowledge about the process. Is it likely that the process is moving the shift or is there a random variation depending upon that, you will apply the test. If it is likely that your data might contain a particular pattern, you will look for them using the respective test. Adding more tests to the jar is not correct. It will only increase the chance of finding a false negative, false positive, or a false signal that the process is out of control. When in reality the process is not out of control. When you increase several test or when you use several tests together, the chances of obtaining signal of lack of control increases. Let's understand how if, in case you're working with variable data, you will be using variable control charts. If you're not sure which test to apply, by default, Minitab always applies the rule number one, the test number one, that is, how many damped the data point is outside the three standard deviation. But apart from that, you might try using test1, test2, and test seven. When you apply this based on the control limits, if they are based from the data. After the control limits are established, you must use the known values for that limit. Then we don't need the test number seven. Number 11 from control limits, detects a single out-of-control point. That's to nine points in a row on one side of the central line detects the possible shift in the process. This seven too many points around one standard deviation of the central line. It detects whether the control limits are too wide and Vida control limits are caused by stratified data, which occurs when you have a systematic source of radiation within each subgroup. Let us understand if you're working with attribute data of defects and defectives. In this case, you are not sure which test to use UGA free to use desk number one and number two. Test number one is about 1 away from the control limit. This number two is nine points and draw on one side of the central line, detecting the possible shift. After the process is established. You are control limits are always drawn based on the values from the data that you have supplied. We will be doing a lot of practical exercises in the next lessons. So stay connected. And if you have any doubts and questions, please feel free to write in the discussion section below, and ensure you try to apply the concepts outside. Please ensure you do your project and upload the project. Do write your review and share your thoughts about how did you feel this class when this number. In the next video, we're going to learn about which test is applicable in my control charts. 12. Which tests are available with my control chart?: Let us understand which test is applicable in my controller. That's number one to eight are available in most of the variable control chart. Note that only test one to four are available in R and S chart and moving range chart, that's number one to four are available inaccurate control charts. Which test is applicable in my control chart? On time better control Jan MiniTab only performs a test for points that go beyond the control limits. That is your number one, our test number one, that the seven test assumes that the points are independent because the plotted points are in time ordered, the weighted towards combine the information of the previous subgroups and the points are not independent. If you have any queries, please feel free to ask in the discussion section below. In the next video, we are going to understand what are the different types of data and how I've worked with them in the control charts. 13. Types of data for control charts: In this video, we are going to understand the different types of data that we have for our control charts. What type of data do I have? This is the question you want to answer. You, if your data is about continuous, determines if you have something like length times B, which is a continuous datatype it from your process. Then you're going to use continuous control charts like IMR chart, Our x-bar R-chart, x-bar S chart. But if you have some multivariate process, then you are going to use multivariate controls. If your data is an attribute like defects and defectives, we're going to use attribute process control charts, which are like n chart, P chart, C chart. There are some charts which are called as a very when process chart, which we will be covering separately. So what type of data do I have? Control charts that you use depends on whether you collect the continuous data or attribute data. If you have multiple content, continuous variable, consider whether you have multivariate data. Mike, continuous variables are infinite numbers such as 84704 something, so you can endlessly divide them. Attribute data have two subtypes, binomial and Poisson. The values of an attribute data are restricted to specific categories are distinct values. For example, attribute data could be like paths and feel. The number of defects and a sample can also be an attribute data which is fallen a Poisson distribution. Continuous measurement usually provides more information than that attribute data. Remember this? However, the attribute datas are generally easier to collect because you just have to accept, but it's defect or an effective and how many defects if it's defective. Does the attribute data is often collected when the continuous measurements are difficult to obtain. Attribute data are often subjective rating that are assigned by the operators and quality control person. Because I feel it's a defect I counted. If I don't feel it's a different, I will not come to that dependency is there? Let us understand the control process data. Continuous data measures the characteristics such as lend, weight, temperature, etc. The data often includes fractional or decimal values. For example, a food manufacturing manufacturer wants to investigate whether the weight of the cereal product is consistent over time. To collect this data, the quality analyst records the wheat from a sample of the cereal boxes. If you have any questions, please feel free to ask in the discussion section. And do ensure that you complete your project and try to apply these concepts. Do write your review. 14. Use Case: Help the Quality Engineer: Let us continue our understanding of control charts. Control charts are also called as statistical process control, SPC. We do this exercise during the control phase of our Six Sigma project. Spc was developed by Dr. Stewart in 1924. He said that you, as the concept goes, y is a function of x. We should not only monitor the project, why, but we should also monitor and control the vital fuels are the x's which are contributing to that y. So by continuously monitoring the x and y together using the control chart, it becomes easy for the project owner and the process owner to monitor the performance and keep it in control. They control processes proactively. It's not. It can clearly identify what are natural causes and water assignable causes. Natural causes are nothing but common cause and assignable causes are nothing but special cause. It also helps you to identify and prevent process from this special causes. If you look at a control chart, this is just a sample. You will have upper control limit and lower control limit, which are created approximately a three standard deviation from the central line, which is usually the process mean. And these three numbers are getting calculated from the process data which you have captured. Anything outside the upper control limit are outside the inner lower control limit are called as special cause variation and their assignable cause. If you're seeing a variation in your process which is entangling or going around the central line. These are due to the common cause variation. This is just a sample representation. The reason he took three standard deviation from the central line is that if you remember, in the bunker, 99.73 per cent of the data is getting covered within plus or minus three standard deviation. So whatever variation you are seeing that is 0.135 on left and 0.135 on the right. That is because of the special cause variation. Now, again, what I covered earlier, I'm just repeating because this is very important and it's also an interview question. What is the difference between control limit and specification? The control limit is getting calculated from the data that book control and the lower control. And it describes the water. What is this process capable of achieving? Specification limits, on the other hand, is given by the customer and management. And they specify what is the process requirement. It describes what the process should achieve to be able to continuously be called as a capable process and it is able to meet the customer requirements. So again, control limits from data specification, limits from the customer. Control limits help you gets calculated and it helps you identify the common cause and special cause variation. If you have data points within control limit, but outside the specification limit, it means that your process is not capable. We will be looking at those examples in the future. When we have this control charts, do I have showed you that it's a plus or minus three standard deviation. But those calculations depend upon the type of data that you have. When you're doing control chart. We have variable data and attribute data. Readable data is also called as continuous data. Where you have decimal points like length, time, distance. These have can be continuously divided. So if your subgroup sizes one, then we go for X IMR chart or individual moving range chart. If your sample size is between 29, you go for X-bar, R-chart, or winter. But if our subgroup sample is greater than ten, it is easier for us to calculate the standard deviation because I have more number of data points in each sample, then I will be drawing an x-bar S chart. On the other hand, if I have an attribute data or discrete data, the attribute data can be of two types. Is it a defective data or is it defects data? What are we monitoring? If it's defective data and the sample size is constant, it could be any number, it will be 1020304023 or 12. But that number, whatever you're taking is constant. Then we go for NP Chart because the control limits are getting calculated based on the formulas from the underlying chalk. The variable sample. For defective data. Like for example, I want to calculate the number of defectives. But today I did a production of 100 units to more sturdy, I did a production of 95 units before that day was 96 units. So my sample size is reading on a daily basis. I will be using the p chart. We will be seeing examples of Peter, where you'll understand that the control limits for a P chart and U chart are zigzag. We will cover that. Do not worry. We are going to have lots of practice so that all these concepts become very easy for you to record. Let's continue. So if I have defects data, It's the sample size is constant, then I use c-chart. If the sample size is reliable, I go for Utah. The thing which I easily try to remember is if it's defective, so it's a piece, the whole piece gets defective. That is why we have P and NP Chart. Because it's constant, we go for NP and because it's variable, it's a p-chart. The other is obviously C. And because C is for constant seated and other other one is Utah. So I tried to use this jingle to remember that which are to be used. So if it's defective, it's P or NP depending upon the sample size. If it's defect that is C and U. And between C and UC is what constant and Q is what variable? Let us take a use case. Though. There's a quality engineer who is at an automotive part. He wants to he wants to monitor that how the length of the cam shafts are getting done. He has three machines which the company uses. They work 24 by seven in three different shifts. So what the engineer does is that he's taken a sample of five from each machine during each shift, you would have easily guessed because the sample length is a continuous data. Sample size is five. It means that it is between 28 and I will be going for X-bar R-chart. So let us understand how are we going to do this. I have already shared with you the datasheet, right? If you come to the main sheet, you have cam shaft length data. I just have to come and click over here. It takes me to that part where the data is present, right? I have given you a lot of data but you do not have to search. Just use the mean sheet for surgery. I have machine 123 and sample IDs. So I'm just going to copy all this data and paste it in my data sheet. So I'm going to take all this data. I have my MiniTab handy. I'm going to paste this data here. Then I'm going to click on Stat control charts, variable with some group. And it is because I know it's a subgroup size of less than eight. I'm going to go for X-bar R-chart. Let the pop-up come. Yeah. So all the observations in the charter in northern column, yes. Our observations of the subgroup are on one. So I'm leaving it like this and I'm going to select Machine 123 and subgroup size is five, right? And I can also use the subgroup ID because I can see I have 111115, right? I'm going to click on, Okay. I will want to really do some mistakes for you to understand what mistakes we need to avoid. Now, it is telling X-bar chart from machine to machine. I click on Output To see all. It has created one chart for each machine. So X-bar chart for machine one. And if you look at over here, you can find one red dot outside the upper and lower control limit. So it's a special cause variation. But in the range chart, everything is in control. Let's scroll down. Let's see, machine to machine to the process appears to be in control, and the range also appears to be in control. If I see the x bar in jail for machine tree, again, I have two data points which are out of control. Now one more thing which you need to observe is the upper control limit. Let me just zoom in a little. The upper control limit for machine one is 6.64301 and lower is fine. 98. Let's go to machine two. It is six hundred and five ninety eight. If I go to machine three is 60298. So can you see that the upper control limit is getting calculated separately because the control limits come from the process data. And there is a variation, the control limit, because there is a variation in the data. Despite the upper control limit being 602, we have points which are going beyond the control limits. Right? And same way, I would request you to look at the control limits for the rain chart between machine one. Machine two has a range of 0 to seven, and machine, sorry, machinery has ranged from 0 to seven. Machine two has a range from 0 to two, and machine one has a range from 0 to five. This gives you an idea that reinforces the concept that your control limits are getting calculated from the process data. I can do the same chart and come to multiple graphs. And I can see the same, but I would say, okay, I can go to X-bar, R-chart, go to the test, and select the test which are important. So if you remember, we said that you should be testing for test1 and test2 to understand it very well. So let's click on, Okay, let's click on Okay, and let's redo this chart. You can see that Minitab is recalculated. If you see the X-bar chart for machine one, it still the same one data point outside the control limit. Now, if you're comparing this, you can see that because I said same by it is very clearly showing that the process is very sharp and very narrow. And CMV four-inch also it is very less. And machine three and machine three R-chart. At the bottom you can very clearly see it says the test reason for x-bar chart in machine one, test one field 1, more than three standard deviation from the central line. Test fail at point number eight. So you can go ahead and identify the subgroup of a where the point number is eight, and then investigate what happened on that day. We learned right? We need to do brainstorming to, to fix the special cause variation. The test results for x-bar chart for machine three, test 11 from three standard deviation from the central line. This test has failed at two places, Point number 2, number 14. So it becomes very helpful for us to investigate what happened on that particular date. We will continue with more examples in the next video. 15. Use Case: Can-filling Process : Let us do our next exercise of helping the quality engineer with the can wait. A quality engineer at accounting company assesses whether the filling process is in control. As you know, these are aerated drinks. If they are more than required, it will cause a leakage or the blast of the box. And if it is very less the customer is dissatisfied. To check if the process He's in control. The engineer collects a subgroup of n gans to minimize the within group variation. That is a can-do can variation within each subgroup. The engineered collects the data of the given subgroup in a short period of time. Now, let us understand what type of chart should I use? What is the process we are taking the variable data because I want to pick how much quantity of liquid is filled up and what are the subgroup size. It is ten. So I need to use an x-bar S chart. Let's get onto our project file. On the main chain, you have the data. Click on it. It will take you to the place where the data is pleased. I'm going to copy this data into Minitab. Now click on Stat control charts. Readable data with subgroup and x-bar S chart as my number of samples in each subgroup is ten, it's more than eight. All the observations are in one column. Yes, so I'm going to put as the subgroup size or the subgroup ID. I'm going to mention it because it's going to be 1 third, I'm just going to click on x-bar s option for both test and I prefer to do test number two as well. Click on Okay, click on okay. Now let me look at the output. So the output that we got is the X-bar chart, where we can see that the test has failed. On point number three, the range has not failed, but the X-bar chart has failed. So test1, 1 out of three, more than three standard deviation from the central line. The test failed at point number three. But asking the quality engineer to go back to the data and see that what happened when he, when when was this data collected? And what happened during this process, that the process was out of control. And as you can see, the process is out of control on the lower. To limit. It means less quantity of liquid was felt. The quality engineer now needs to bring strong that what happened during that hour. And is it something that's an assignable cause? And how can we avoid that special cost from happening? Again? Let's take up the next example in the next video. 16. Use Case:Injection Molding Process: Now let us help the other quality engineer from a plastic company. If you can see this, it's a plastic molding process. A small animation which is blue by VR, is shown over here. So you can see that the plastic mold comes over here. It goes through the channel where there is lot of heat. Hence the plastic gets melted. It gets into the molding section where the food is kept and when the product comes out, it's a nice toy which we can sell in the market. So the use case is the quality engineer for a plastic part company monitors and injection molding process. The machine has a dye that creates £5 in one time and they engineer collects 20 subgroups of five-part each. They engineer monitors both within subgroup variation and between group variation at the same time. N between the batches. So as we're trying to monitor the plastic injection molding process, we have a subgroup size of five and He's selecting 20 subgroup data points. So obviously, what's the data that we are going to look at? What type of chart do we need to work with. My sample size is less than eight, so I need to use an X-bar R-chart. It's obviously the variable data. Let's go to our control charts. I have my injection molding data over here. I'm going to copy the data of the pods and the subgroup into MiniTab. Now, this time I'm going to show you this process using the assistant feature in MiniTab. So when you click on Assistant, go to control charts. It's continuous data. Collected data in subgroup Yes. Is the subgroup less than eight? Yes. So I go for an x-bar in charge. You can see how intuitive it is over here to work with many tapped. My data is present in part. It's not a constant size, So I haven't seen the column WhatsApp group ID is over here. How do I determine the control limits and central line? I'm saying estimate it from the data points. It has immediately identified that there are certain points where the data is missing out on x-bar and R-chart. So I leave as it is and click on. Okay. When I look at the output, my x-bar R-chart four parts has been created. Is the process being is a process that we are monitoring staple. It says, no, the process is not saving. The process that we're monitoring is not stable. The subgroups are out of control in the X-bar chart. Keep in mind that you may see a 0.7% of out-of-control subgroups by chance. But more than that is called as a special cause variation. When we look at this chart, this time, we have a problem in the X-bar R-chart. The previous subgroup was missing the point on the upper control limit. And the next subgroup is missing the point below the lower control limit. It is definitely required for the quality engineer to investigate what happened during the point number 13 and the sample that was collected during point number 14. We also find that the ring chart is also having an out-of-control limit. Which means that the range, the sample that was collected during the point number eight had a wide variety of radiation. So was it an assignable cause or is there a bias? When the other data is collected? The engineer needs to investigate it. Whenever we're doing an X-bar R-chart, we look for certain patterns. Is there a global trend like the way you're seeing over here? Do you see some cyclical pattern? Currently in this data? I'm not seeing any global trend. I'm not seeing any cyclical pattern. Is there a shift in the process? I can see that, yes. There is a slight so the point was down and then there is a ship which is going up. Is there adrift, like it was completely down and up. I don't see any drift over here. Is that an oscillation of data? Not at the moment. Can you see some mixture patterns? I can see a little bit over here. Can you see some excessive out-of-control? I can see it both in my x-bar and R-chart. So it says very clearly the X-bar chart has missed the control limit on by number 1314. The R-chart went out of control during point number eight, which makes it easy for us to understand that. Why should we go and investigate, right? You'd need not be concerned about the precision of the control chart limits because 70% or more data points are included in the calculation. Your data has passed a correlation test. The correlation between conjugative data points within each subgroup is less than 0.02. It means that the randomness is there when the sample was collected. But yes, we have found that the process is out of control, which is requesting the quality engineer to investigate into the matter. I hope you are enjoying the practical exercises of the control chart. I would also request you to practice all this from the data sheet and upload it in the project section. If you have any questions, please feel free to ask in the Q&A section. We will continue with one more example in the next video. 17. Use Case Detergent pH data: Let us help the quality engineer from a detergent company. This is a use case where the quality engineer wants to establish the pH data. The quality engineer wants to map and monitor the manufacturer of the liquid detergent and wants to assess whether the process is in control. Engineer measures the pH or clarify conjugate two batches of detergent. And because the data is not collected in subgroup, you are required to use the IMR chart. He's using 25 constitutive batches and ask what he's measuring is the pH, which is a continuous data. We are going to use an IMR chart. I'm going to show you one more time our types of control charts because this will help you remember it for life. In control chart, the type of data we are measuring his pH, which is a variable datatype. Each batch he's selecting one sample value and that's why we, n is equal to one. And hence we are going to go with the IMR chart. Let us go to the project file. This is the project data file that I have sent our uploaded in the discourse section. Scroll down, you will find the detergent pH data. Click on it. It will take you to the place where the data is present. I'm now going to copy this data into Minitab. I have coped. I have copied this data into Minitab. I need to now build my IMR chart. There are two ways. One, I can go to stats, candle jogs, and then go to weird charts for individual and click on IMR. But this time let us use the assistance. So I'm going to do assistant, click on control chart. Data type is continuous, is the subgroup data collected. Know I'm going to use the IMR chart. The data volume is Beard. And I'm want to estimate the control limits and the central line from the data. It is very clearly told the point number three. There is a possible point which is out-of-control. As a quality engineer, I never delete any data point because that's a point for me to investigate. I just click on Okay and come to see my output. Yes. The IMR chart for peace is this process means table. It says, no, the process mean is not stable for 4% of time. The process may not be stable for per cent of the data points are out-of-control in the eye chart. Keep in mind that 0.7% of out-of-control is by chance, even if the process is stable. But now I have more data which is out of control and this data point has missed on test number one, it might rain moving range chart. My process is not out of control. We look for certain patterns whenever we are building our control charts, like trends, cyclical shift, drifts, oscillation mixture, and excessive points out of control. In my current process, I can very clearly see that there is an out-of-control point. I'm not seeing mixture on oscillations, shifts and drift in my current control chapter. The process variation is Steven, no point is out of the control limit in the moving range chart, but there is 1 which is out of control in the individual charter, the I check. If the data is not normal, you can see that this could be a false alarm rate. So you have to also do the normality test, which has been covered in the other series. As a quality engineer, the advice I would give to this engineer is to ensure that the quality gets monitor and investigate what happened during the point number eight. I hope you understood the concept and you will be able to apply this in your own project using your own data. Apart from that, I would request you to practice using the dataset that is given over here. 18. Use Case Steel bar length data: Let us move on to the next use case in control charts. Here, let us help the quality engineer who wants to determine whether the steel bar cutting process is in control. They engineer measures the length of five steel bars, some tents, ships. Can you guess what is the datatype that we are using? And which type of control chart should we be using for determining if the process is in control? Can you help the quality engineer? Can you type in the Q&A section that what type of date control charts should we be using? Thank you for being engaged. You are right. As the control chart is available data, that is the length of the state bar and my subgroup size is between 2528, that is, my current subgroup size is five. I will be going ahead with the X-bar R-chart. Let us take the project file. I have my data project file, which I have already shared with you in this project. So you should go and pick up steel bar land data. Click on it. It will take you to the place where the data is present. I'm going to copy this data into Minitab. Yes, the data is present in Minitab. I have copied the data from my datasheet. Let us do the analysis. I click on Stat control charts, variable that subgroup x-bar in charge. I have placed length in the data column and subgroup ID in the subgroup size. I go to the x-bar or option and go to the test section and ensure that the test number 12 are selected. You remember we learned this during the earliest stage. Which test to use? When I click on Okay, I click on Okay. Minitab is going to do the analysis and get me out. If you see this, it has prepared the X-bar chart for us. I can very clearly see that there are no data points which are going out of control. Whatever variation we are seeing in the process is due to some common cause variation, we need to continue to monitor this process. As I can see that there is a data point touching almost the upper control limit. I haven't request the quality engineer to collect some more data to be sure that the data processes in control. I would request you to do the similar exercise, create your project and upload the project in the project section. I'll be happy to review your project and give you the feedback. And this will give me confidence that you are all enjoying what you're learning and you are also able to apply what you're learning. Let us continue with another example in the next class. 19. Use Case Unanswered calls data: Hello friends. I guess you are enjoying and learning a lot from this control chart. Lessons. We have seen examples of x-bar, R-chart, X-bar chart, and IMR chart. Till now, let's get into the attribute type of data. Here we have an example from the call center. The supervisor of a call center wants to determine whether the call answering process is in control. The supervisor records a total number of incoming calls and the number of unanswered goals for 21 days. As you know, if we are the customer and we're calling up bunny for a query and our phone doesn't get answered. We feel frustrated. And then we don't want to go back to that company to work again, to partner with them or buy the products from them. Hence, Unanswered call is a major problem in the contact center industry. And we have to help the supervisor to understand how he can reduce it and whether currently is his process in control or not? Because the type of data is count of defectives, the entire goal is not answered, so it's not defect, but it's defective. And do you can you control the number of incoming calls on a daily basis? They are variable. Hence, we need to use the character. Datatype is attribute because it's the number of calls. And my data types sub datatype is defective because I don't answer a part of the colon, so right. Either answer the color, I do not answer the call. Each column is a piece. The total number of calls received in a day is a variable number. Hence it will be a variable sample size, and hence we need to go with the p-chart. Let us go and see our data in the Excel sheet. Does the product data file that I have already shared with you? Click on an answer. Calls. Data is present over here, and I can very clearly see that on few days, I have one twenty three fifty three calls to 65 calls to 58 calls, and so on. I have the number of calls that are not answered. So I would go ahead and copy this data into Minitab. I have copied the data into Minitab. Now I need to perform the test. I can go ahead, click on Stat control charts. The datatype is attribute chart, and I already know that I need to build a p-chart. I click on it. I see unanswered calls and the subgroup size is total calls. I go to the P chart option, go to the test, and I can determine which all tests do I want to test. So I'm going to click on all the four tests to see if any of these desktop failure. I click on Okay, I click on Okay, my data is produced. Let's view the output. We can see the p-chart for amounts are called. And my data is very much within the control limit. And it has not violated any of the four tests that we perform on the p-chart. One interesting thing which you might have observed is that the lines are zigzag. The control limit is not the straight line as the way you saw in the X-bar R-chart or the IMR chart. Can you guess the reason why? You are right? Because my sample size is wearing, my control limits are also vary accordingly. And hence the p chart would have is exactly the test is performed and there is no radiation. The process is in control. The variation is due to common cause variation. If you have done the same exercise, I would request you to copy this graph, save it, and upload it in the project section of this course. I'll be happy to review your project and we'll be happy to share my feedback. If you have other data relating to defectives, you want me to review it. Please upload it as a project. You are not going to violate any NDA because you're not going to share the data. You're already going to share the chart with me. Please ensure you do not upload any of your company data onto Skillshare. You'll just upload that information where you need clarification without revealing the name of the client or the company from where you are uploading. Thank you. We will learn more in the next session. 20. Use Case Defective Umbrella P chart: Let us do one more example of the featured. In real life. Most of the time we are into manufacturing and production. Even if we are in the service industry, we consider the client is always monitoring as on defects and defectives. And hence, I'm showing you more and more examples for the different attribute type of control charts. So this is again, one more place where the supervisor of an umbrella manufacturing unit wants to evaluate the quality of production. As you know that it's a monsoon season now in India, sale of umbrella is at its peak. So if the manufacturing setup does not produce good-quality umbrella, they even not be able to sell it. And the remaining umbrellas will stay back with them as a stock, which they will be only able to be selling the next financial year or the next month soon. So to avoid that, the supervisor wants to record the total number of umbrellas that were produced are manufactured every day and the number of defectives for 21 days in a series. Because 21 is a good number, we can go ahead with this. Again, I repeat as we are checking for defective datas and the sample size is reliable. I'm going for the pitcher. Let's go and copy the data from our data sheet into Minitab, which I have already done. I go to the main sheet, I have umbrella data. Let me just scroll up umbrella data. And I have copied this data from here to the MiniTab. I go to assistant, I go to control chart, I go to the p-chart and C number of defective columns. So it's defective umbrella that is constant subgroup size. Know the column of subgroup size is total produce. I want to estimate it from the data and I click on Okay. Easter defective umbrella be charged for defective umbrellas. Is the proportion of defective items table? Yes, it is stable. The proportion of defective items table, there is no subgroup, which is why leaving the room. You would see when I'm doing the p-chart using the assistant, my control limits are coming out as a straight line instead of the zigzag line. That is one problem with the p-chart. If I'm using an assistant, if I would have done the same using stats, control charts, attribute data, and p-chart. Taken the defective. And to reproduce what are the p-chart option, go to the test and ensure that all the desk directive click on Okay, click on okay. Now you can see that it's showing the p-chart for umbrella as a zigzag line. Right? Perfect. However, because the line is very far off, it's okay. Even if it was a straight line because you are not missing out on any point. Unless we saw what we saw earlier where we had the call when we had this type of zigzag lines. And hence here it was very important for us to use the normal. We'll go into stats control chart and doing it. Wonderful. I'm happy that you have been practicing with me lots of case studies on learning it a lot. If you have any questions, please feel free to write in the question and answer section or the discussion section below this. I'll be happy to answer your queries and be happy to help you out with any doubts that you have. Let us continue for other examples in the next video. 21. Use Case Hospital Medical Records Defects data: Let us continue with the next example in the control charts. This is an example of a use case from the hospital. The hospital maintains the medical records of the patient. As a quality engineer, there is a problem that we're facing. The supervisor of a small hospital wants to ensure that the number of arrows in the hospital medical records remain in control because it's dealing with the life of the patient. Supervisor records, the total number of medical records that were filled each day, and the number of records that are incomplete or inaccurate that is defective. So as you know that the total number of records that we spend each day is a variable number and we're talking about a defective record. So can you guess what is the type of chart we need to prepare? Yes, you are right. We need to prepare the p-chart. So because the data is related to the counts of defectives, we're going to use the picture. Again to recap. Our data is attribute data, defective variable sample size. Hence, we are going to use beta. The reason I keep showing you this again and again is to ensure that this gets imprinted in your mind. Now let's go to our project file. In the project file, you have to go and look for the data which is about medical records. So it can you see it's defective medical records. I will click on this and it will take me to the place when the defective record is there. I'm going to copy this data. And we can see very clearly that the sample sizes wearing and the defectives are also different. So these are the total number of records, these other defectives. I'm going to copy this in mind. Minitab. Yes, I have the data here. As we are going to use the P chart, I can go to stat control charts, attribute charts, and select the p-chart. Total number of the variable is defectives, and the subgroup size is the total number of records. I would go to P chart and go to the test and ensure that all the protests are active. I click on Okay, I click on Open. The output has come out. Let's go to the output window and look for what has happened. We can see that there are multiple places where the P chart has on test number one. It is point number 810 thirty five, fifty six seventy five eighty seven eighty nine. We need to go to our data and look what happened on that day that we have so many of the factors. The defectives are on the below the lower control limit and as well as the upper control limit. Though, this is a positive change that we have making less defectives. But we need to understand how can they be so careful and why can't we reinforce the same pattern all the time so that our control limits will change. You can also see that the data, the control limits are highly zigzag because the available data size is wearing. Each day my sample size is wearing. And based on that, mice control limits are highly zigzag. Okay, so now I hope you understood what needs to be done. You need to investigate into the matter. On the subject matter experts go to the point number eight and investigate what happened on that day. So I would come here and go to point number eight and see that, Oh, out of 1700 records, 1778 records only 3D vectors. It's a positive thing. But why can't I repeat this behavior again and again on the other this, if it was possible to have some good data on the particular dates, I would want to reinforce and repeat the good behavior. I hope you understand. I would request you to practice creative project file. Save this project file as an image and ensure that you upload it. In the project section. I have created a separate video which tells how can I upload the project. I would be loving to review your project and give you the feedback and share the experience that you have gotten by learning to my class. I will continue the rest in the next video. 22. Use Case Defective light bulbs data: Let's take one more example about defective data. Here we have the light bulb data. Let us help the quality engineer in this company. The quality engineer assesses whether the process used to manufacture light bulbs is in control or not. As you know, the bulk can be either completely defective or it is it puts on the light. It cannot be half defective. Hence, it's a defective data and not differenced data. The engineer test 500 light bulbs each hour for 38 hour shift. As the sample size is constant, we are going to use the defective data. For constant sample size. The Engineer records the number of bulbs that did not light. Hence, he's recording the defective bulbs. As you understood. We are counting the number of defectives with the constant sample size. We are going to use the np chart. Let's take the recap of our types of control charts. We're looking at attribute data, we're looking at defective data, and we're looking at a constant sample size. Let us now look at the dataset that we have. Look for this bulb data. I'm just Yes. So defective light bulb data. I have two fields over this group ID and defectives the number of light bulbs that did not lie within the subgroup. The data is over here. I'm going to copy this data into Minitab. The subgroup Id, which you look see over here is an identifier, but the size of the subgroup is saying 500. Let's go to MiniTab. And I have pasted my data over here. As it is a constant sample size. For defective data, I'm going to click on Stat. Control charts. Attribute data, np dot. I'm selecting the vectors and the subgroup size is 500. I go to NP Chart. Click on Test and ensure that all the four marks are ticked. Click on Okay. Click on Okay. Now let us reflect on the control chart. We can see that at this point, the test number three has failed. And at this point the test number one has failed. So what is the test number 11, more than three standard deviation from the central line. This test has failed at point number 16. There's 36 points in a row, all increasing or decreasing. This has happened on point number nine. If you see from point number 32, number nine, we have a continuous decreasing trend. So we would want to investigate what was happening during this six ships that the number of light bulbs which were effective, was constantly reducing. Because it's a positive behavior. And we want to reinforce this positive behavior. If the data would have been completely shifted and it will be between 02, our control limits would have changed. Hence, as a quality engineer, you're not only looking at something which is outside upper control limit, but you might also want to reflect on the behavior of the data that can avoid things going out of control. Good. So I hope you understood this. If you have any questions, do not forget to ask your questions in the discussion section below, I'll be happy to answer them. I will continue with the next example in the next video. Till then, happy practicing and happy learning. 23. Use Case Wallpaper defects data: Now we will move to one more example. Here we will try to help the quality engineer in a wallpaper manufacturing company. As you know, post COVID, most of the families have decided to renovate the house because they were at home for the last two years. And they realize that their housemates redevelopment. So there's a lot of demand for the wallpaper. And hence it's important for the quality engineer to manufacture good-quality wallpapers. But one purpose, as you know, come with various designs. And Prince, he wants to understand that he's the printing process table or not. Three hour, the engineer takes a sample of 100 feet of wallpaper and count the number of printing defects, such like pattern distortion and ink missing. As we are looking at the different data with a constant sample size. Can you guess what type of control chart should we be building? Your right? I need to build for the constant sample size for different data. Again, I'll take you through the types of control charts. So we're looking at attribute data defects and a constant sample size. So we're going to take up seizure. Let's go to our project file. Look for the wallpaper defect data. I can find it over here. I click on it. It takes me to the place where the wallpaper data is present. Here the sample ID is only for identifying the id entity should counting the number of defects, each sample is of a 100 feet, right? I'm going to copy this data into Minitab. I have done the same. You remember I need to do with jot, correct? I click on Stat control charts, attribute charts, and C for constant sample size. I'm going to measure the defects. I'm going to go to C chart and ensure that all my tests are selected. I click on Okay, I click on Okay. Now let's look at the output. There is the output of the c-chart. As the sample size is constant, my upper and lower control limit are straight lines. I can see that the test one has failed at two points by number 12 and boy number 13. If I do not want to count at which point it is, I can look at the reference which MiniTab also the test has failed at point number 2, number 30, it investigates a matter that why did so many defects happen on these days? While there was so much of high printing errors that were encountered during these two shifts. And again, due to what it came back in control. So it should definitely be an assignable cause which needs fixing. Good. If you have any questions, do not forget to ask your questions in the discussion section below, I'll be happy to answer all your queries and don't forget to complete your project and upload it in the project section. I would invite you to do multiple projects so that you are confident of what you are learning. And this gives me a lot of confidence, even as a facilitator, that I have been helpful in your learning journey. Thank you. 24. Use Case Defect Medication errors: Let us take one more example. As the pandemic just got over, it's important for us to check if all the hospitals are working under control. This is a quantity in Junior who is checking whether the records in the medicines or the medicines are being given at the correct rate and other character. So the Director of the quality for a group of hospital wants to ***** the medication error rate. Example of errors include delivering medications at the wrong timing, delivering the wrong dosage, or delivering the wrong medication altogether. The director records the number of patients and the number of medication error each week for 32 weeks. As you can understand, the number of patients will be variable. And we are referring to the defects data like wrong dose, age, wrong timing, drone delivery of medication. We are going to use the Utah. Let me take you through the types of control charts. It is the attribute data, it is the defect data readable sample size. Hence, I'm going to use to your chart. Let's go to our project data file. Let's search for medication error. It's heal arrows, the number of medication errors that happened that week. Patients at the total number of patients each week. Mentioned over here. So when I go over here, I can see how many errors and how many patients they can very clearly see that the patient's size is reading. I will copy this data into Minitab. The data is in Minitab. Now, let's do the analysis. Click on Stat, click on control charts, click on attribute charts, click on Utah. Arrows are over here. The subgroup size is the patient's. I go to the U chart, click on the test and ensure that all the fence for desktop selected. I click on, Okay, I click on Okay. My output is present. Let's pull it up. Now. We can see that there are multiple points at which the test one has filled. Based on the data, the upper control limit and the lower control limit is being calculated. We can find that either there are very less arrows are, there are very high error. So in, on both the sides, the three standard deviation is getting this note. So if there are consistently, actually this has a positive effect. If you are getting less errors for the number of patients, it's a positive effort. So as a quantity engineered, the dots at the bottom will be a point of investigation that how can we manage it so well? Whereas the doubt dots on the top beyond upper control limits are the places of concern that why have we missed and meet so many arrows? And did it cost us the life of the patient? So as the quality person for this group of hospitals, it's extremely critical for you to investigate in this matter because the test has failed at multiple points. And do remember, control limits are not specification limit. As per the specification limit, the group of hospitals might have agreed that at 10% errors is acceptable or 7% errors is acceptable. So all the points beyond that will make your process less capable. I have covered that in the other video, which talks about the hypothesis. Let's continue focusing only on control charts in this chapter. Before I go further and getting you back to my project file, you can see that there are multiple examples which are given over here. I would invite you to practice them because this will give you confidence of continuing and understanding the concepts. We have covered multiple examples till now. Let me take you right. We saw that medication error example. We saw the wallpaper example. We saw the defective light bulb example. We saw that umbrella example. We saw the hospital medical record example. We saw the contact center example. We saw the detergent p-hat example. We saw the length of steel bar. We saw the injection molding. So what does it mean? I have already taken you through multiple examples in each of the chart that I would invite you to practice all of it because this will give you confidence of how do we investigate. The motto is my process in control. What investigation do I need to do? With that? I will stop over here and I'm looking forward to your queries in the discussion section. And your project completed, uploaded in the project section. Happy learning and continue to grow. Thank you. See you in the next lesson. 25. Thank you Note for my students: I thank you very much for completing this series on control charts. You would have practiced all the examples that I showed you in this lesson. You will have also factors that using the data that I have already shared with you, you should have, you should also try to practice control charts using your own data, which will give you a different confidence level. I dressed, you have learned a lot and hope your concepts are also very clear. Please follow me on Skillshare. This is my profile and I will keep uploading new videos as we go forward. For those who are interested in doing some corporate training. I do hi interactive corporate training where I do programs. I have workbooks which is specially designed for my participants. Depending upon the company where they're working. You can see that how everybody is engaged in this photograph. These are just a glimpse of some of the trainings that I have done. I have done many virtual programs during the pandemic, which accounts to more than thousand hours of training. These are just snippets from some of them. For those who would like to stay in touch with me. You can connect with me on LinkedIn. I have a Telegram channel which is called as six underscore. Six, number six underscore sick. Well, you can connect with me over there as well. My e-mail ID is also displayed on the screen. If you have any queries, please feel free to contact me. If you have any questions. Do not forget to use the discussion section given below. I'm looking forward to your learning process and I hope I can be of help in your learning journey. Thank you so much.