Statistics Article

Adam Conner-Simons’s picture

By: Adam Conner-Simons

This story was originally published by MIT Computer Science & Artificial Intelligence Lab (CSAIL).

Scatterplots. You may not know them by name, but if you spend more than 10 minutes online, you’ll find them everywhere.

They’re popular in news articles, in the data science community, and perhaps most crucially, for internet memes about the digestive quality of pancakes.

Jay Arthur—The KnowWare Man’s picture

By: Jay Arthur—The KnowWare Man

There are two ways to increase profits: increase sales or reduce costs. Although most data analysis seeks to find more ways to sell more stuff to more people, addressing preventable problems is an often overlooked opportunity. Preventable problems consume a third or more of corporate expenses and profits.

Data analysis can pinpoint problems and eliminate them forever. Problem solving with data is a much more reliable and controllable way to cut costs and increase profits. Sadly, few people know how to do this consistently.

How do you solve operational problems with 100-percent success rate? Take out the guesswork. The vast majority of improvement projects involve reducing or eliminating defects, mistakes, and errors. If you have raw data about when the defect occurred, where it happened, and what type of defect it was, you can create a world-class improvement project that eliminates the guesswork. And you can do it using a tool you most likely already have: Microsoft Excel.

Matthew Bundy’s picture

By: Matthew Bundy

Untitled Document


Burning plastic cart carrying a fax machine, a laptop computer, and a three-ring binder. Click here for larger image. Credit: FCD/NIST

Several centuries ago, scientists discovered oxygen while experimenting with combustion and flames. One scientist called it “fire air.” Today, at the National Institute of Standards and Technology (NIST), we continue to measure oxygen to study the behavior of fires.

Douglas Allen’s picture

By: Douglas Allen

Any number derived from real observation is made up of three components. The first of these is the intended signal, the “perfect” value from the object being observed. The second is error (or noise) caused by environmental disturbance and/or interference. The third is bias, a regular and consistent deviation from the perfect value.

O = S + N + B, or observation equals signal plus noise plus bias

The signal usually is predictably constant, as is the bias. Identifying and eliminating bias requires a set of techniques beyond the scope of this article, so for the remainder of this, we will consider both as components of the signal, leaving a somewhat simpler equation for our observation.

O = S + N, or observation equals signal plus noise

This article focuses on removing the random noise component from the observation and leaving the signal component. The noise is in the form of chance variation, which sometimes enhances the signal and sometimes detracts from it. If we could separate the noise from the signal and eliminate it, our observation would be pure signal, or a precise and consistent value.

Anthony D. Burns’s picture

By: Anthony D. Burns

Augmented reality (AR) means adding objects, animations, or information, that don’t really exist, to the real world. The idea is that the real world is augmented (or overlaid) with computer-generated material—ideally for some useful purpose.

Augmented reality has been around for about 30 years. But it’s only during the last five years or so that it has been widely used on mobile devices. If you have wondered why your new iPhone 12 has a LiDAR depth sensor, the answer is, in part, for augmented reality. Almost all modern phones now have depth sensors for AR. LiDAR makes depth sensing more accurate.

Unlike virtual reality (VR), AR on mobiles requires no special equipment. There’s no need for headsets or handheld devices. All you need is your mobile phone.

More than fun and games

Although games are probably the most notable use of AR on mobiles (Pokémon Go is a good example), there are business and training applications as well. Perhaps the simplest AR business application is labeling real-world objects. Google Maps, for example, recently launched Live View, adding real-world labeling of objects and directions via the mobile phone’s camera. Real-world objects, when viewed through the mobile phone, can show added text, objects, or 3D animations. Live View has all of these.

Donald J. Wheeler’s picture

By: Donald J. Wheeler

On Sept. 29, 2020, the recorded worldwide death toll from Covid-19 reached 1 million. Six days earlier the United States reached 200,000 Covid-related deaths. So how did the United States with only 4 percent of the world’s population manage to capture 20 percent of the world’s deaths in this pandemic?

The 19 countries listed in figure 1 account for 85 percent of the Covid-related deaths worldwide, as reported by the European CDC. Here we can see how the U.S. death toll exceeds all others.

 
Figure 1: Number of Covid-related deaths reported by 19 countries as of Sept. 26, 2020

The short explanation for this dubious achievement is that between April 1 and the present, the United States had an average of 27 percent of the worldwide total number of confirmed cases of Covid-19. With that kind of market share, the high death toll was sure to follow. But a more detailed answer requires that we look at the number of deaths per capita and the rate at which these death tolls are growing.

Eric Weisbrod’s picture

By: Eric Weisbrod

The idea of digital transformation can be scary. The growth of technology is outpacing a comfortable pace of adoption for many manufacturers. But remaining content with the status quo often means being left behind. Digital transformation has become an imperative to give manufacturing organizations the flexibility and agility required to overcome business disruptions and adapt to rapidly changing and demanding global markets.

Digital transformation of quality management is a process that depends on something you already have: quality data. Your quality management system is key to optimizing all your quality operations, including supplier and materials management, production processes, quality checks, packaging, and shipping.

InfinityQS calls this holistic approach “manufacturing optimization.” It starts with improving the way you use data to answer the strategic, big-picture questions that truly matter to your business.

Limits of the status quo

The barriers to transformation are often a result of operational and resource challenges that typically boil down to one thing: everyone’s plate is already full. Whether managing and maintaining servers and IT projects, or running day-to-day production, no one has the time to take on new transformation projects.

Steve Wise’s picture

By: Steve Wise

The importance of data analysis in manufacturing operations can’t be overstated. Over the years, manufacturers have used statistical process control (SPC) methods and tools to study historical data and reveal differences between comparable items: shifts, products, machines, processes, plants, lot codes, and more.

The foundational benefit of statistical methods is predicting future behavior from historical data. That’s why control charts, box-and-whisker plots, Pareto charts, and the like are so valuable: They indicate that if processes are not changed, then performance (positive or negative) will continue as it is.

Control charts are brilliant tools for assessing performance over time, and their related “control limits” are predictions of normal future behavior. The problem is that many SPC software products struggle to move beyond just data collection to offer truly insightful data analysis.

Multiple Authors
By: Dirk Dusharme @ Quality Digest, Jason Chester

In previous articles of this series, we discussed how to master quality at the tactical and strategic levels. If you are like most readers, you probably nodded your head through article two’s tactical shop-floor view and vigorously shook your head through article three’s strategic view because your organization has the same challenges.

There is understandable hesitation from nearly any organization to make the transition to ultimately master quality at the enterprise level. This hesitation stems from organizations trying to view this transition as an all-or-nothing endeavor. As a result, that gets people seeing roadblocks that don’t necessarily exist.

Let’s take a look at a few.

Solve all the deployment issues in advance

Because previous deployments took a tremendous amount of time, resources, and expense, most organizations want to address every deployment problem before they even begin considering strategy.

That won’t happen. Ever. No matter what system you deploy, you will discover and learn things you could never have anticipated up front.

Jason Chester’s picture

By: Jason Chester

Before we get into a case study about how enterprisewide SPC software would work on both the shop floor and the C-suite, let’s talk about a long-held bias about “blue-collar” workers: That because they’ve traditionally been associated with manual labor, they should use manual tools; “white-collar” front-office workers, on the other hand, need the slick technology tools.

Imagine walking around the offices of a large manufacturing organization and finding salespeople managing customers’ information using a Rolodex. In a planning meeting, the CEO is using acetates on an overhead projector. In the procurement office, staff are issuing purchase orders using a Telex machine.

Now imagine walking the plant floor at that same manufacturer. The production supervisor is writing machine settings for the next shift on a board next to the machine. The quality engineer is writing the results of a critical quality check on a clipboard with a blunt pencil. A bunch of people stand around murmuring, scratching their heads, and wondering why a machine isn’t working properly.

In the first example, you might think you’d traveled back in time. The scenes are absurd. But the second example is a common reality.

Syndicate content