All entry results tagged with “data visualization”

Today is the first of two guest blog posts this week. Aaron Crews graciously agreed to write this while I'm goofing off this week on vacation. Thanks Aaron!

Jim is a brave soul to hand over the keys to the blog while he's on vacation this week. Hopefully I don't do too much damage while he's gone. If I do, maybe Mike Boudreaux can help me roll back the odometer before Jim returns.

As you might remember from some previous posts on this blog, I am an engineer with The Automation Group (TAG), which is part of the Emerson Process Management family. We provide management and technical services for all types of process control system projects. I thought I would take this opportunity to share a story from one of my past projects.

Here at TAG, one major focus of our work is on DCS modernization projects. These projects arise for a variety of reasons, usually including system maintenance issues, a desire to take advantage of new technologies, and to improve overall plant performance.

A big part of this last goal depends on the operators. The importance of data visualization in the operator interface, alarm management, and human-centered design within the HMI have all been hot topics for that reason. In a modernization project, however, that is only part of the equation.

The operator is required to make a big adjustment from the old system to the new one. Graphics may look totally different, navigation between screens is done in a different way, alarm notification and acknowledgement isn't the same as it used to be, and even the way in which the operator physically interacts with the control system can be different.

Every project includes an operator training component, but as anyone who has been through training knows, it's different once you're on your own. That is one reason why we usually recommend a hot cutover between the old and new systems - so that operations gets some transition time and experience before the critical parts of the process are cut over.

On this particular project, however, the trepidation from plant operations was especially strong. In migrating from RS3 to DeltaV, they were changing from a dedicated keyboard with a trackball and keys with particular functions that they had memorized to a traditional keyboard and mouse. After going through some training and getting their hands on the new system, the operators began to feel like it just wasn't going to work for them. Pretty soon, a dedicated keyboard had become a make or break issue for the project.

It was hard to disagree with how they felt. As a "power user" of several pieces of software, I know I can work dramatically faster and more productively when I utilize keyboard shortcuts. Eventually everyone will get used to the new system, but the process doesn't stop and wait for operators to work their way back up to speed.

As an engineer, though, I appreciate a challenge, and I was able to come up with a solution that would please the operators, mitigate the safety and production risks associated with the modernization, and put the project back on track. Using the flexibility of the DeltaV system and its built-in key macro functionality, my project team customized a commercial off-the-shelf, physical, programmable keyboard for DeltaV.

The top half of the keyboard was dedicated to display navigation, giving the operator single-button access to any graphic. The bottom half of the keyboard had a layout designed to mimic that of their RS3 keyboards, with analogous functions assigned to each key.

Needless to say, the keyboards were a hit. They didn't require custom software on the DeltaV stations, they didn't require specialized knowledge to maintain or replace, and the fixed-button layout allowed for quick "muscle-memory" reaction to the process while always keeping the operator focused on what he is doing and not how to do it. The keyboards were not required so eventually they could be removed if the operators decided that they prefer the regular keyboard and mouse.

Half of the 5000+ I/O was cut over hot, and the rest was cut over during a short turnaround, and the plant was brought back up without incident. I have visited the plant several times since then as they have gone through a couple of DeltaV upgrades and the keyboards are still going strong. They are enjoying all the advantages of DeltaV and experiencing none of the worries that they had with their previous generation control system.

Thanks again to Jim for the invitation to contribute here. If you'd like to learn more about what we are up to at TAG or if you have any questions, you can email me, follow me on twitter, or find us on Facebook.

Marshall Meier received an award for his recent Emerson Exchange presentation, Improve How You Visualize Data. I first met Marshall at last year's Exchange after I discovered we both did presentations on social media, a.k.a. Web 2.0. Here was last year's summary of his presentation.

In this year's data visualization presentation, he shared ideas from preeminent thought leader, Edward Tufte, and his seminal book, The Visual Display of Quantitative Information. Another thought leader Marshall highlighted was Stephen Few, who wrote several books on data presentation, including Show Me the Numbers.

When presenting, your focus should be on just showing the data. Don't clutter it with extraneous visual information. Why does this matter? Your data is important. It helps you and those with whom you share this data better understand problems and make decisions. The goal for an excellent graph or chart should be to get the viewers of this data to ask questions.

Marshall showed the same information presented in both a cluttered and clean presentation. The first view might prompt more thoughts about mad scientists than about the rise and fall of U.S. patents over time. Presentation viewers are more likely to engage in a conversation about the data with the simple uncluttered presentation.

One of Tufte's ideas Marshall shared was the data-to-ink ratio. This is the amount of ink devoted to the data compared to the total ink used in the graphic. For a PowerPoint presentation, think pixels instead of ink. The higher the ratio, the more you're focusing the visual display on the data.

Marshall described "chartjunk" as additional graphics not related to the data in a quest to make the chart more aesthetically pleasing. Instead, it distracts from the data. If you think your graph is boring, you're showing the wrong data. Open a USA Today newspaper and you'll find examples of chartjunk--like a gas price tracker that shows a gas pump graphic with a window containing the average price and another containing the directional price change. Although cute, the data-to-ink ratio is low.

Another example Marshall shared is the comparison between the main Yahoo page and Google page. They are at opposite ends of the visual display spectrum.

For those that love the visually exciting, 3D graphics in Microsoft Office products like Word, Excel, and PowerPoint, Marshall's simple message is, "Don't use 3D." Although it may look cool, the 3^rd dimension does not add anything. When the graph has multiple data series, 3D makes it difficult to compare data between two data points. In 2D, comparisons are easier to make. Again, the focus shifts from the "eye candy" to the data itself.

You can also fall into the trap of distorting the data. The concept of "lie factor" is the size of the effect shown in the graphic divided by the size the effect in the data itself. Graphic treatments that show a 3D-perspective can visually distort the data. Bar charts often convey size, relative differences, and area better than pie charts do.

Marshall closed his presentation with a famous graph described by Edward Tufte:

Probably the best statistical graphic ever drawn, this map by Charles Joseph Minard portrays the losses suffered by Napoleon's army in the Russian campaign of 1812.

It shows you can use creativity and "eye-candy" if it amplifies the data and not the extraneous. Congratulations, Marshall, on a very informative, award-winning presentation!

MP3 | iTunes