The Clear View with Enact® Bubble Charts

Introducing new insights and decision-making power into your quality data with bubble charts. The bubble charts in our award-winning Enact® cloud-native Quality Intelligence platform are as elegant and easy to interpret as they are powerful tools.

Bubble charts are not new. Far from it. They’re just new to manufacturing quality software. Like the scatter plot, a bubble chart is primarily used to show relationships between two numeric variables. However, the bubble chart adds even more functionality. Combining different-sized bubbles and time-based animation with the x and y axis plotting on a standard scatter plot provides four dimensions of data that can be incredibly valuable.

Important Questions

The bubble charts we use in Enact® offer manufacturers the ability to view quality data in a new, multi-dimensional way. These charts are useful in many ways, but particularly the ability to distill thousands of data streams into a visual that provides answers to the two most basic quality data questions:

  1. Are my required checks being collected on time?
  2. Are my processes and products meeting standards?

Context

Manufacturers are, and should be, particularly concerned with data collection. Manufacturers need to know that data are collected on time, and that the processes and products are meeting standards. They also need to be able to see how these two critical metrics might be changing over time. Are yields going up or down? Are the percentages of on-time checks getting better or worse? How does one process compare to others? Are there certain areas where yields are high, but on-time checks are low? Which areas are showing the most improvements? Or, degrading?  The answers to these questions will shed insights into what needs to be fixed and where to discover best practices.

A bubble chart is a data illustration tool that can display four dimensions on a two-dimensional graph. The basic components are the aforementioned x and y axes, as well as magnitude and time. Visually, it’s a collection of circles (bubbles), with the area of each circle proportional to the number of measurements. This “organic” appearance can be intriguing and intuitive and, in the case of their use in Enact®, extremely useful because their visual feedback is immediate.

Figure 1: The X and Y axes represent the variables to be studied. The TIME axis, when activated, will illustrate how the X and Y change over time. The bubble size represents the amount of data represented in the bubble.

Ultimately, we use bubble charts in Enact® to help manufacturers make intelligent, efficient business decisions—by enabling them to visualize huge data sets. By looking at large data sets, the user can isolate patterns of interest and process behaviors over time.

But, to be clear, we realize that the bubble chart is not the be-all or end-all. It is not “everybody’s everything” analysis tool. Taken in its context, and utilizing it for its strength, the bubble chart is a shortcut that can help you avoid poring over stacks of charts or histograms…and it can lead you to the correct streams of data that require further interrogation—via tools like Box & Whisker charts. The bubble chart, as it turns out, can help you—the manufacturer—reduce variation in your manufacturing processes. And isn’t that what this is all about? Isn’t that why you collect data to begin with?

Reducing Variation

So, why did we add this function to a tool like Enact®, which is already great for helping you with what we call “dysfunctional” data? Colleague Jason Chester, Director of Global Channel Programs for InfinityQS, reminds us that “The underlying causes of performance and quality challenges in manufacturing today often can be traced back to a single problem—data.” In order to achieve manufacturing excellence, to continually improve your manufacturing processes, there is no way around the fact that you must collect reliable data and you must collect it on time.

Then, when that data are no longer dysfunctional—when they are complete, consistent, stored in a single repository, and efficient—then, and only then, can you interrogate your data to obtain valuable insights into your manufacturing processes; insights that will lead you to strategic decision making that can transform your business. Bubble charts can help you see into your data and lead you on the path to your organization’s transformation.

Bubble Chart Setup

Unlike traditional quality tools like a control chart, which limits the user to a single part/process/feature stream of data, the bubble chart monitors key process indicators across thousands of data streams and rolls them up to the department, site, region, etc. to provide a high-level view of performance. This high-level view enables users to quickly expose areas in the company that need attention (as well as those areas that are doing well) and then drill down to the individual processes, if desired.

Because the bubble charts can be “played” over time, one can also isolate areas that are showing improvements, as well as those areas that are degrading.

There are three types of metrics that can be configured as either the x or y axis: sampling compliance, yield, and number of events.

Sampling Compliance
Sampling compliance metrics tell you how timed data collections are performing. For any timed data collections, the sampling compliance can be viewed in terms of:

  • % Completed
  • % On-Time
  • % Late
  • % Missed

Yield
Yield metrics are a great indicator, as they directly relate to the performance of your processes and manufactured products. The yield can be calculated from either:

  • % Expected Yield
  • % Measured Yield

The measured yield is based solely on the fallout from the values entered into the system. Expected yield is a projection calculated from the mean and standard deviation of the entered values.

Number of Events
Event counts are all about compliance. The event type is selected to further isolate root causes and corrective action options. The types of events that can be selected are:

  • # Out of Specification
  • # Above Upper Specification Limit
  • # Below Lower Specification Limit
  • # Net Content Control Violations
  • # of values that are greater than T2 (upper)
  • # of values that are less than T2 (lower)
  • # of values that are greater than MAV (upper)
  • # of values that are less than MAV (lower)

Getting Started with Bubble Charts

There are many metric combinations you can configure for your bubble chart, but we recommend using % On-Time Checks, as this metric is easy to understand and is an intuitive “bigger is better” metric that goes hand in hand with % Expected Yield, which is also a “bigger is better” metric.

Figure 2: In this figure, each bubble represents a manufacturing site. The bright green site contains the most data, but that site’s performance is only 30% On-Time Checks with 72% Expected Yield. Whereas the little blue site is doing much better at 65% On-Time Checks with a 98% Expected Yield.

These two metrics are a great starting point to expose trends that will lead into more focused questions and stream analysis. Another eye-opening paring is % Expected Yield vs. % Measured Yield. The results of this analysis could expose systemic problem areas where the entered values do not reflect reality.

Figure 3: Notice that for most sites, the % Measured Yield is better than the % Expected Yield. For example, the red site’s % Measured Yield is 97.25% whereas the % Expected Yield is 82%. This points out that the reported results (% Measured Yield) do not represent the expected results.

Bubble Size Statistic

Optionally, the user can configure the bubble sizes to reflect the number of measurements contained in that bubble. Bigger bubbles contain more data points and therefore will appear bigger compared to bubbles containing less data.

Reading a Bubble Chart

Manufacturing supervisors and managers often have responsibility for multiple lines, or even sites. Their time is valuable, and they usually need (or want) to simply see “how things are going.” Quickly. Visually. Bubble charts can help immensely with that need.

Enact® takes advantage of a centralized data repository, using bubble charts to display data visually and intuitively for easy comparison. When the chart is “played,” and bubbles begin to move over time, you can easily spot trends, outliers, and problems with your data collection efforts.

This information can lead you to a number of conclusions. Is there a dramatic difference in on-time checks between plants or lines? Do operators on a certain line or in a particular plant require additional training? Are the right notifications taking place and going to the right people? Management can then take this information and suggest corrective actions to improve data collection operations.

Tool of Choice

You can see how useful bubble charts can be. By allowing manufacturers to visualize huge data sets, it’s easier for them to make intelligent, efficient business decisions. By looking at those large data sets, supervisors and managers can isolate patterns of interest and process behaviors over time. This, of course, is very useful for making sure your organization’s data are collected on time, and that the values are believable and reflect true process and product yields.

Bubble charts are not a panacea, nor the be-all or end-all, but you need to know the relationship between yield and on-time checks so you can see how things are changing, or not changing, as your products are manufactured. And Enact® bubble charts are a great tool for doing just that.

That’s peace of mind. And peace of mind is like your own personal bubble of protection.

.

By Steve Wise
Vice President of Statistical Methods, InfinityQS
July, 14 2020