EigenU Europe returns to Copenhagen, DENMARK October 17-20, 2022 Complete Info Here!

Category Archives: Uncategorized

Python is free

Apr 5, 2022

PLS_Toolbox is not free. 

But you don’t have to be a dedicated data scientist to use PLS_Toolbox (or its stand-alone equivalent Solo). Many of its users are, but the real expertise of most users is in something else such as analytical instrumentation (typically spectroscopy) or the specific problem they are working on (e.g. chemical process control, disease detection, art provenance etc.). We made PLS_Toolbox because we believe that the best data analysts are the people that generated the data and have the physics, chemistry or engineering background to understand it.

PLS_Toolbox includes a very wide array of tools for pattern recognition, data visualization, sample classification and regression plus many data preprocessing tools. Tools for problems specific to spectroscopy like calibration transfer, curve resolution and variable selection. Plus particle analysis, batch modeling tools and tools for reading all sorts of data files from various analytical instruments. It is pretty much one stop shopping for most people that work with analytical chemistry and related data. 

Unlike Python, when you use PLS_Toolbox you don’t have to decide first which of the 40 most popular libraries you need. It doesn’t require a nine page cheat sheet. You can use it from the command line if you want, and script it too, but the vast majority of analyses can be done using the highly refined point-and-click interfaces. And when you are comparing model results from different methods, you can be sure that they are evaluated in precisely the same way, apples to apples. 

And if you can’t figure out how to use a tool or think you’ve found a bug? There’s one email address to write to: helpdesk@eigenvector.com. We have five full time equivalents working on it, and one of them will get right back to you with help that’s actually helpful. We’ve been supporting it for more than 30 years (and we have no intention of stopping). Plus we have other data scientists on staff who can help you with your application when you really get in over your head. 

Yeah, Python is free. We like that about it too. That’s why we search through Python libraries to find the tools that PLS_Toolbox users will find useful. We then incorporate them with our wrappers and interfaces around them so they behave the way our users have come to expect. It’s why we say “we learned Python so you don’t have to.” 

So what’s your time worth? If you are someone who proudly displays Dr. in front of your name or Ph.D. after it, you are worth at least a couple hundred bucks an hour. Those hours of command line bullshittery add up pretty fast. Not to mention the opportunity cost of not being focused on the problem you’re actually trying to solve. 

So yes, PLS_Toolbox is not free. But for many if not most analytical scientists it is a better value proposition than Python alone. 

BMW

Remembering Svante

Jan 6, 2022

Svante Wold passed away on January 4, 2022, at the age of 81. In the fields of chemometrics and data science, there is no need to use his last name. If you say “Svante,” everybody knows who you are talking about.

I first met Svante almost 30 years ago at the Chemometrics in Analytical Chemistry conference in Montreal, CAC-92. I was a pretty new Ph.D. at that point, having graduated the year before. I presented a talk, “Monitoring the Health of Multi-Channel Analytical Instruments with Multivariate Statistical Process Control (MSPC).” Before the meeting, Bruce Kowalski told me to be sure to talk to Svante. So at some point, perhaps after his talk, I walked up and introduced myself. The first thing Svante said to me was something to the effect of “We’ve been following your work on MSPC and find it very interesting.”

I was shocked. I had no idea such an influential professor would even want to talk to me, let alone have an interest in the work I’d been doing. But that was Svante. Always interested, positive and easy to talk to. Very encouraging to young people in the field, especially those who were working with latent variable/projection methods.

Attending many of the same conferences we ran into each other frequently after that. I remember especially fondly the Gordon Research Conference (GRC) on Statistics in Chemistry and Chemical Engineering. I attended my first GRC in 1993. Roger Hoerl was chair that year with Svante as his vice-chair, who would become chair the following year. Much to my surprise, Svante organized a side meeting to discuss development of PLS_Toolbox, which I had started distributing freely a couple years previously. The next thing I knew, “young Barry” (Svante’s nickname for me) had been elected vice-chair of the GRC. This meant that I would work with Svante in ’94 and become chair in ’95 (with Age Smilde as vice-chair).

The GRCs were exceedingly interesting scientifically and great fun. There was ample time for discussion, technical and otherwise. It was there that I discovered what a hoot it was to have beers with Svante. Many jokes and funny stories. (I still tell the Digger joke.) Svante always had a crowd around him. This of course continued at many more conferences and other gatherings over the years.

Svante had some unique views on data modeling. For instance, he once told me that he had found it useful to be somewhat lazy when modeling, and that philosophy had always served him well. “Don’t try too hard,” he said. This is good advice: the harder you dig and the more tightly you fit a model, the more likely you are to get a spurious correlation that won’t hold up. If what you are looking for isn’t readily apparent at first, you should be very cautious!

Along with my wife, Jill, I was very lucky to get to spend time with Svante and Nouna at their homes in Umeå, Boston and Hollis. We enjoyed their hospitality greatly. Unfortunately, we were never successful in talking them into coming out West to see us.

Barry M. Wise, Svante Wold, and Jill Wise at EAS in 2013

I’ll close with a picture that for me captures a bit of what was Svante. It was taken at Eastern Analytical Symposium (EAS) in 2013. We were all just goofing and having a good time at the President’s Ball. You can do a literature search and easily determine Svante’s vast influence on chemometrics and data science. But when I think of Svante I remember the fun times like this.

Rest in peace Svante, you will be long remembered and sorely missed.

BMW

New Year, New Courses and Webinars

Jan 12, 2021

In response to the COVID-19 pandemic, in 2020 we offered many online courses, over 100 hours of lectures in total (not counting the classes we did for specific clients). We gave ourselves a break towards the end of the year but we’re now re-energized and ready to get going again.

Our first webinar of the year is scheduled for January 21, EVRIthing You Need to Know About Logistic Regression. Our Manuel Palacios will describe this important classification technique which is used in machine learning, medical applications, social sciences and many other fields. Logistic regression was added in the recently released PLS_Toolbox and Solo Version 8.9 and of course Manny will show you how to use it!

On February 16 Basic Chemometrics PLUS Online commences. This instructor-led live short course includes 7 classes drawn from our Eigenvector University series. The “Basic” part covers linear algebra, Principal Components Analysis (PCA) and common regression methods including Partial Least Squares (PLS). Together these classes provide an onramp to machine learning in the chemical data sciences.

The “PLUS” part includes 4 classes that are making their online debut:

Model Maintenance Road MapI’d like to draw special attention to the Calibration Model Maintenance course. It’s an unfortunate fact that most multivariate data-derived models do not last forever. There are many things that can change (e.g. instrument behavior, feed stream compositions, etc.) and cause model outputs to be biased or noisy. Worse still, in many instances no plan has been developed to deal with this reality. In Calibration Model Maintenance we’ll go over our Roadmap (at right) and also cover methods and tools (such as instrument standardization/calibration transfer methods and methods to detect model output drift) that can be used as part of an overall model maintenance strategy.

Courses on CLS and MCR are new online offerings but have been mainstays in our live Eigenvector University syllabus. The Robust Methods course, offered less frequently, covers methods to automatically identify and reject outliers so that models can be built on the consensus of the data. These methods are especially useful in data sets with large numbers of outliers or “bad data,” typical of chemical process data.

Hope to see you at one of our webinars or courses soon. Stay healthy!

BMW

Year 27

Jan 5, 2021

And so it begins again!

For the 27th time I’ve gotten up on the first business day of January looking forward to a new year with Eigenvector Research. The overall plan doesn’t change much: keep improving the software, provide training in data science/chemometric methods and help clients with their challenging data problems. But the details! So many things to decide! What features to add to our software? What courses to teach, and when? What research to pursue and where to present and publish? The list goes on…

Plus we welcome this week a new staff member, Sean Roginski (pictured at right). Sean completed his M.S. in Data Science at University of North Carolina Wilmington last fall. Prior to that Sean did an internship with us developing a Python software development kit (SDK) to enhance communication with our Solo_Predictor prediction engine. He also constructed an architecture that will help us integrate Python tools into PLS_Toolbox. We really liked what he did, so we hired him! And yes, if his name sounds familiar, he is the son of our Bob Roginski.

The addition of Sean Roginski brings our full time software development and support staff to five, including (in reverse order of seniority) Lyle Lawrence (3 years), Donal O’Sullivan (11 years), Bob Roginski (14 years) and R. Scott Koch (17 years). Given their 50 man-years of experience with our software (Bob and Lyle were PLS_Toolbox users long before they became employees) it’s no surprise that they make our Helpdesk actually helpful.

So we look forward to bringing you new tools and courses in 2021 and, as always, are available to help you with your data science projects.

Stay healthy and stay tuned!

BMW

LCVSF 2020 Scholarships Announced

Aug 4, 2020

The Lake Chelan Valley Scholarship Fund has announced 15 award recipients for 2020. The scholarships go to college bound seniors from Chelan Valley schools including Chelan High School and Manson High School, and also renewals for previous award recipients. This year’s crop includes three graduates from this year’s MHS class: Brenda J. Alonso Arreola, Nadia Tejada and Bryce LaMar; five from CHS: Casey Simpson, Tobin Wier, Eli Phelps, Emma McLaren, and Kyle Jackson; and seven renewals: Colt Corrigan, Henry Elsner, Jessica Oules, Jasmine Negrete, Joe Strecker, Sierra Rothlisberger, and Quinn Stamps.

Each recipient will receive $2500 for a total of $37,500 in awards this year. COVID restrictions permitting, checks will be presented to the recipients at the flagpole at Chelan Riverwalk Park at 9am, Saturday, August 8.

LCVSF Board President Betsy Kronschnabel noted “As the years go by, I continue to be impressed with how the majority of applicants are focused on their future plans and have already established clear pathways to their goals. We are really happy to be able to help them move forward.” 

The LCVSF was made possible by Doug and Eva Dewar, who wished that their estates be used to help the children of the Chelan Valley. LCVSF was founded in 1991, and in that year five scholarships in the amount of $1000 each were awarded. The fund has grown substantially over the years from contributions from many people, but especially significant contributions from John Gladney, Ray Bumgardner, Don & Betty Schmitten, Marion McFadden, Virginia Husted, the Dick Slaugenhaupt Memorial and Irma Keeney.  Now in its 30th year, LCVSF has awarded over $625,000 to Chelan Valley students since its inception.

LCVSF accepts applications from residents of the Chelan valley for undergraduate education. The awards are renewable for up to four years. LCVSF welcomes applications from graduating high school seniors as well as current college students and adults returning to school.

The LCVSF board includes Betsy Kronschnabel (President), Arthur Campbell, III, Linda Mayer (Secretary), Sue Clouse, Barry M. Wise, Ph.D. and John Pleyte, M.D. (Treasurer). For further information, please contact Barry Wise at bmw@eigenvector.com.

Tags: lakechelanscholarships

How to Run a Webinar or Online Course

Jul 8, 2020

We started running “how to” webinars with our chemometrics software in the Fall of 2019. When COVID hit we decided to step up the frequency of these, and to add short courses. Up to now we’ve done ten webinars and five short courses running from 2-5 days each. Both have been quite popular. We’ve had as many as 200 people online for the webinars and 80 for some of our classes.

Having had a positive experience, (over 98% of respondents to our post class surveys say they are likely or highly likely to recommend our online courses), one of our short course attendees asked if I would provide our advice on running these online events. I was happy to do so, then realized that there are others out there that might benefit from our experience. So here it is!

Participant’s On-screen View of a Webinar

Eigenvector Research has never had a central office. We’ve all worked from home for 25+ years now. We started using WebEx to talk to each other long ago, so when we had to choose a platform for webinars we stuck with it. We’ve had good luck with WebEx as a platform. It is maybe a little old and clunky compared to Zoom, (and we wish they’d stop changing their interface), but it is definitely industrial strength and has not been plagued with security issues like Zoom. We don’t do Zoom at Eigenvector. 

We record our webinars and classes (more on that below) but we’ve found that people really prefer to attend the live sessions. Besides the possibility of asking questions, there’s just a better connection to the speaker and the content when you’re watching live.

For all our online events we have three key positions, the host, the moderator and the speaker(s). The host is responsible for starting the meeting, controlling who is the current presenter, and introducing the speaker and moderator.

With big groups it really doesn’t work to have people ask questions verbally; it creates too much chaos. We have participants ask questions through the chat feature and the moderator monitors the chat. The moderator then decides what questions to pass to the speaker and when would be a good time to interrupt. Usually, we have the moderator go over these “rules” when he is first introduced.

The speaker or speakers, of course, deliver the main content of the webinar. Having the moderator filter the chat questions for the speaker works well. For shorter talks questions could be held to the end, but interruptions are actually good. When you speak too long on a WebEx without feedback you begin to feel pretty disconnected. So relevant interruptions are welcome. And for these audience, it makes things a lot less monotone to have some back and forth between the moderator and speaker.

I would advise speakers against trying to monitor the chat as they talk. It’s too distracting, better to have the moderator do that. And it’s really great if collaborators of the talk being presented are online at the same time to help answer questions through the chat. There are usually several Eigenvectorians in attendance that can answer questions in the chat that really don’t need to go to the speaker, such as questions specific to the person asking and not that useful for all the participants. And they often provide links to supplementary material that can be quite useful.

Before any online event be sure to practice with the roles and make sure everyone is on board with procedures. It is especially good to practice how to pass presenter privileges and how to activate screen sharing mode. Also, test the audio! We’ve found that headsets work best. Even cheap ones are way better than using your computer microphone.

Speakers should always present their slides in “Presentation” mode so they fill the screen. It’s important that things be as large as possible when transmitted because they get shrunk into a smaller window on the participant’s end (see photo above). And definitely have speakers make up the slides in “Widescreen” format. If you use the old format you get wasted black space on either side of the presentation. If they are going to do anything on their own screen, e.g. software demos, they should change their screen resolution to something lower than native, like 1600 x 900, so everything on it is bigger. Otherwise it can hardly be seen on the participant’s end. 

When demoing software, be sure to go slow. Explain what you are doing in words. Hover over buttons while dithering the cursor before clicking. Rest momentarily on fly-out menus before selecting. We’ve been pleased to find that about 75% of the participants that attempt to follow along with our software demos at home are able to do so, but we are working to make that even better!

In WebEx you should have the “Mute on Entry” turned on and “Anyone can share” turned off. Even so you’ll find that people will get their microphones turned back on so you need to have the host or somebody with host privileges ready to track these people down and mute them. 

We use video sparingly because it eats up people’s bandwidth. I would suggest that the moderator and speakers turn theirs on right at the beginning just to let people see there are live people on the other end, but then turn it off once they get going. 

Unless you are prepared for a big bill, when preparing the WebEx meeting invitation make sure you turn off the setting that gives out the toll-free number. That gets billed back to the WebEx account holder and if you have a lot of international’s participants it adds up fast. We got a $1200 phone bill after doing our first Basic Chemometrics PLUS Metabolomics class. Ouch!

We use the features in WebEx to record the sessions. These good quality but they include all the windows around the speakers screen. So we edit them and crop off the surrounding material before posting them. We use iMovie but there are many other options. It takes a while for WebEx to make the recordings available, in our experience up to 8 hours, so in the end it takes about a day to get them downloaded, edited and posted online.

We tried YouTube and Vimeo for our videos and definitely preferred Vimeo. It’s pretty easy to create “Showcases” of related videos in Vimeo, and to password protect them. (We record all our courses and make the videos available to participants in a password protected Showcase.) Vimeo also provides good analytics on viewership and it’s easy to embed the videos in web pages. For some examples, see our Webinar Series page.  

We often provide supplementary videos that are posted with the recorded live events. We use Snagit to record our non-live videos and have had really good luck with it. There are many other options of course. 

The chat sessions can be saved and provided along with the recorded videos but they may have to be edited. The saved chat from the host includes all the messages that were sent privately between the host and other individuals. But sometimes there is some valuable stuff in them. It is hard to remember to save them, however, and in WebEx they are very hard to retrieve after the fact!  

Of course the MOST important key to a successful online event is to have content that people actually want to see. We’ll leave it to you to figure that out! Good luck with your online endeavors!

BMW

Sociopaths, Enablers and Due Diligence

Apr 15, 2020

I just finished reading “Bad Blood: Secrets and Lies in a Silicon Valley Startup” by John Carreyrou. It is the story of Theranos and its founder Elizabeth Holmes. Our Bob Roginski sent it to me after I mentioned that I’d seen the HBO documentary “The Inventor: Out for Blood in Silicon Valley.” I highly recommend the book over the much shorter documentary.

I’ve followed this story a bit since first hearing of Theranos and their claims to be able to run hundreds of blood tests simultaneously on just a few drops of blood. Based on my experience this seemed more than unlikely. At Eigenvector we’ve worked on quite a few medical device development projects. This includes projects involving atherosclerotic plaques, cervical cancer, muscle tissue oxygenation, burn wound healing, limb ischemia, non-invasive glucose monitoring, non-invasive blood alcohol estimation and numerous other projects involving blood and urine tests. So we’ve developed an appreciation of how hard it is to develop new analytical techniques on biological samples. Beyond that, we’ve also learned a lot about the error in the reference methods we were trying to match. Even under ideal conditions, with standard laboratory equipment and large sample volumes, results are far from perfect.

So when the whole thing was blown wide open by Carreyrou’s reports in the Wall Street Journal I wasn’t surprised. I read several of the follow up articles as well. But as one reviewer of the book said, “No matter how bad you think the Theranos story was, you’ll learn that the reality was actually far worse.” I’ll say. Honestly it took me a while to get into the book, in fact, I put it down for a month because it just made me so mad.

We’ve had a few consulting clients over the years that were, let’s say, overly enthusiastic. To varying degrees some of them have been unrealistic about the robustness of their technology and have failed to address problems that could potentially impact accuracy. (I’m happy to report that none of our current consulting clients fall into this category.) In some instances things we saw as potential show stoppers were simply declared non-problems. In other cases people abused the data including cherry picking and grossly overfit and non-validated models. (My favorite line was when one client’s lawyer told me I didn’t know how to use my own software.) We have had falling outs with some of these folks when our analysis didn’t support their contentions.

But none of the people we’ve dealt with approached the level of overselling their technology to the degree that Holmes took it. As I see it there are two reasons for this. The first is that Holmes is a sociopath. Carreyrou said he would leave it to others to make that assessment but it seems obvious to me. Maybe she didn’t start out that way, but it’s clear that very early on she started believing her own bullshit. Defending that belief became all that mattered. And she teamed with Ramesh “Sunny” Balwani who was if anything worse. They ran an organization that was based on secrecy, lies and intimidation. And they made sure that nobody on their board had the scientific background to question the feasibility of what they were claiming they’d do.

But the second reason they got as far as they did was because they were exceedingly well connected. The book identifies these connections but doesn’t really discuss them in terms of being enablers of the scam that ensued. It started with Elizabeth’s parents connections to people around Stanford, and Elizabeth’s ChemE professor at Stanford, Channing Robertson. These lead to funding and legal help. From there Holmes just played leapfrog with these connections ending at former secretary of State George Shultz, (who I learned actually lives on the Stanford campus) and his circle including Henry Kissinger, James Mattis and high profile lawyer David Boies. (Boies led the Justice Department’s anti-trust suit against Microsoft, and was Al Gore’s lawyer in he 2000 election.) Famous for his scorched earth tactics, Boies and his firm kept the lid on things at Theranos by threatening lawsuits against potential whistleblowers and further intimidating them by hiring private eyes to surveil them. Having no firmer grasp of the science and engineering realities of what Theranos was attempting than the board members, Boies’ firm did this in exchange for stock options.

It’s second point here that really bothers me. There will always be people like Holmes that are willing to ignore the damage that they may do to others while pursuing wealth or fame and fortune. But this behavior was enabled by well-heeled and well-connected people that failed completely on their due diligence obligations from financial, scientific and most importantly, medical ethics perspectives. Somehow they completely forgot Carl Sagan’s adage “extraordinary claims require extraordinary evidence.” It’s hard to imagine that this could have ever gotten so far out of hand had Holmes attended a state university and had unconnected parents. Investors to whom you are not connected and who are not so wealthy as to be able to afford to lose a lot of money have a much higher standard of proof.

In our consulting capacity at Eigenvector we always try to be optimistic about what’s possible, and we do our best to help clients achieve success. But we never pull our punches with regards to the limitations of the technology we’re working with and the models we develop based on the data it produces. Theranos produced millions of inaccurate blood tests that were eventually vacated. While it doesn’t appear that anybody actually died because of these inaccurate tests, they certainly caused a lot of anxiety, lost time and expense among the customers. It’s our pledge that we will always do our due diligence, and expect those around us to do the same, so that Eigenvector will never be part of a fiasco like this.

BMW

UW ChemE Graduates Barry M. Wise and Neal B. Gallagher Celebrate 25 Years at Eigenvector Research

Jan 3, 2020

Eigenvector Research, Inc. (EVRI) was founded by UW Chemical Engineering graduates Barry M. Wise and Neal B. Gallagher on January 1, 1995 and celebrated its 25th anniversary. EVRI is largely based on data science methods in chemistry, i.e. chemometrics, that Wise researched during his graduate work under the supervision of Professors N. Lawrence Ricker and the late Bruce R. Kowalski in Chemistry. Wise developed the first version of Eigenvector’s flagship PLS_Toolbox software as part of his dissertation “Adapting Multivariate Analysis for Monitoring and Modeling Chemical Processes.” The software has evolved over the last three decades into a multi-platform suite of products for building and applying multivariate and machine learning methods for pattern recognition, calibration and classification in the process environment. It’s active user base consists of several thousand people world wide and is used in applications from chemical process control to teaching data science in universities. 

Eigenvector provides training on the use of data modeling methods, and has organized hundreds of short courses attended by thousands of students over the last 25 years. It’s annual Eigenvector University in Seattle, now in its 15th year, is a week-long event typically attended by about 50 students. EVRI also provides consulting services and has worked on a wide array of applications from medical devices to the manufacture of high tech machine parts.  

Wise and Gallagher started their graduate studies at University of Washington in Autumn 1985. They met the first day of class on September 30. An undergraduate in Chemical Engineering and Engineering Physics at University of Colorado, Gallagher completed his M.S. ChemE at UW in 1987. He went on to University of Arizona where he finished his ChemE Ph.D. in 1992. Wise obtained B.S. degrees in ChemE and Chemistry at UW in 1982 and went on to work for Pacific Northwest National Laboratories (PNNL) in Richland, WA. Returning to UW in 1985, Wise completed his M.S. ChemE in 1987 under Professor Harold Hager and his Ph.D. in 1991. 

The two senior members of EVRI’s software development team also have ties to UW: R. Scott Koch graduated with his B.A. in Chemistry in 1995 and Donal O’Sullivan obtained his Ph.D. in Atmospheric Sciences in 1986. They have been with Eigenvector for 17 and 11 years now, respectively, and are a critical part of the Eigenvector team. 

For more background and history about Eigenvector Research please see Eigenvector Turns 25. For questions please contact Eigenvector President Barry M. Wise.

Everybody’s Welcome in the Eigenvector Sandbox

Nov 20, 2019

When I was a kid I was lucky to have a great sandbox. The fact that it was on the shores of Lake Chelan made it especially nice, but there were a lot of other things that made it great. First off, it was big enough for several kids to play in at the same time. It was well equipped: There were buckets and shovels and rakes. And Tonka trucks. And later Matchbox cars. There was a handy supply of water (the lake) and plenty of “building materials” besides sand, including clay that we’d get on the beach, driftwood, sandstone bricks (left over from the construction of my parent’s house) and a few random pieces of pipe. Every couple years Dad would get out the orchard tractor and scoop up a fresh load of beach sand and put it in the sandbox. Each birthday produced new toys. 

Of course the neighborhood kids (about half of whom were my cousins) were all welcome. And it didn’t matter if they wanted to build their own castles or roads or use the toy cars and trucks with the ones built by others. Everybody could play however they wanted. 

Eigenvector’s software suite is a lot like that sandbox. There are a lot of tools (toys)! We’ve got everything from PCA to XGBoost. If you’re the builder type then you can use our PLS_Toolbox and MIA_Toolbox with MATLAB® to script and automate analyses and build new tools. Or you can use our existing highly refined point and click interfaces for a wide array of analyses. Have colleagues that don’t have access to MATLAB? That’s OK, they can use our Solo and Solo+MIA stand-alone software and share data and models. And they can all choose what operating system they like, Windows, MacOS or Linux. 

Unlike the sandbox, where it was typically hard to take one of your creations home with you, our Solo_Predictor and Model_Exporter tools allow users to apply their models outside the sandbox in a wide variety of ways.

So everybody, or as we like to say around here, EVRIbody is welcome in the Eigenvector sandbox. And our developers will keep providing new toys and make sure that the sand stays fresh!

BMW

Ready for MATLAB 2019b

Sep 25, 2019

The MathWorks just released MATLAB® 2019b, and as always, we’re ready! The latest versions of our PLS_Toolbox, MIA_Toolbox and Model_Exporter are all set to go. So whether you are doing PLS calibrations on NIR spectra, curve resolution of pharmaceutical tablets for content uniformity, or SVM models to classify LIBS spectra you can switch to the new MATLAB and keep right on working.

PLS_Toolbox and MIA_Toolbox running in MATLAB 2019b.

As far as compatibility updates go this was a fairly easy one: the changes from MATLAB 2019a to 2019b didn’t break any of our code. We have an extensive suite of test codes that we run to verify this, plus several of us here at Eigenvector always use the MATLAB pre-release in order to identify any issues.

But it’s not always the case that MATLAB upgrades are easy. Sometimes MATLAB updates are significant enough that it requires hundreds of man-hours on our part to make our software work properly. This is especially challenging because we try to maintain a five year window of compatibility with MATLAB versions. The current version of PLS_Toolbox (8.7.1) is compatible with MATLAB 2014a through 2019b.

What does that mean for users? It means that you can stick with your favorite version of MATLAB over the last five years (yes, we all have our favorite versions!) and still use our latest tools. And when you’re ready to switch to the newest version of MATLAB, that will work too! And of course you can also choose between operating systems as we support Windows, MacOS and Linux.

Happy Computing!

BMW

MATLAB is a registered trademark of The MathWorks, Inc.

LCVSF Awards 16 Scholarships for 2019

Aug 11, 2019

The Lake Chelan Valley Scholarship Foundation awarded 16 scholarships to college bound students from Manson and Chelan in a ceremony at Chelan’s Riverwalk Park on August 17. The awardees included four recent Chelan graduates; Sarah Brownfield, Jasmin Negrete, Owen Oules and Quinn Stamps; four recent Manson graduates; Bryan Bernardo, Megan Clausen, Santiago Santana Gonzales and Tyler Charlton; and eight renewing college students; Neil Carleton, Benjamin Charlton, Ahimelec Diaz, Malena Evig, Anabeth Morales Garcia, Henry Elsner, Addie Ivory and Jessica Oules.

This year’s winners will receive $2500 with the exception of one half-year award for $1250 for a total $38,750.

LCVSF also administers the Dick Slaugenhaupt Outstanding Junior Award, which is selected by the Chelan High School Students and Faculty. College bound CHS graduate Mario Gonzales received $1000.

LCVSF 2019 Scholarship Recipients and Board Members

Regarding this year’s scholarship applications, LCVSF Chairperson Betsy Kronschnabel noted “We continue to see bright young college bound applicants in the Chelan/Manson Valley who need a financial boost to get to school.  We carefully evaluate each application based on need, grades and goals, as is the mission of the LCVSF.” Board member Barry Wise added “We’re pleased to be a part of the process that helps support these ambitious young people.”

The LCVSF was made possible by Doug and Eva Dewar, who wished that their estates be used to help the children of the Chelan Valley. LCVSF was founded in 1991, and in that year five scholarships in the amount of $1000 each were awarded. The fund has grown substantially over the years from contributions from many people, but especially significant contributions from John Gladney, Ray Bumgardner, Don & Betty Schmitten, Marion McFadden, Virginia Husted, the Dick Slaugenhaupt Memorial and Irma Keeney.  Since 2004, LCVSF has awarded over $575,000 to Chelan Valley students.

LCVSF accepts applications from residents of the Chelan valley for undergraduate education. The awards are renewable for up to four years. LCVSF welcomes applications from graduating high school seniors as well as current college students and adults returning to school.

The LCVSF board includes Betsy Kronschnabel (President), Arthur Campbell, III, Linda Mayer (Secretary), Sue Clouse, Barry M. Wise, Ph.D. and John Pleyte, M.D. (Treasurer). For further information, please contact Barry Wise at bmw@eigenvector.com.

Eigenvector President Wise Receives Wold Medal

Jun 29, 2019

Eigenvector Reseach President and PLS_Toolbox creator Dr. Barry M. Wise was recognized for his achievements in the field of chemometrics* at the 16th Scandinavian Symposium on Chemometrics (SSC16) in Oslo, Norway. Wise received the Herman Wold Medal in gold “For his pioneering contributions in Process Chemometrics and his extensive, deep commitment to the proliferation of Chemometrics.” The award is sponsored by the Chemometrics Division of the Swedish Chemical Society and was presented by previous Wold medal winner Dr. Johan Trygg of Sartorius, AG.

Dr. Wise is the first American and the first non-academic to receive the Wold medal.

Wise gladly accepted the award and thanked the award committee and his fellow Eigenvectorians past and present for their support. He also acknowledged his good fortune and gratitude for guidance provided by his graduate advisor in Chemical Engineering Prof. N. Lawrence (Larry) Ricker and Prof. Bruce R. Kowalski in Chemistry.

Dr. Johan Trygg (left) presents Dr. Barry Wise with the 14th Herman Wold Medal in gold at the 16th Scandinavian Symposium on Chemometris in Oslo, June 18, 2019.

The text presented at the award ceremony is included below.

Motivation for Barry M. Wise

Dr. Wise has a PhD in Chemical Engineering from University of Washington, Seattle, USA and has been active in chemometrics since 1985. His research has focused largely on chemometrics in chemical process analysis, monitoring and control. E.g. he demonstrated PCA and PLS for monitoring systems to detect process upsets and failed sensors and introduced the term MSPC (multivariate statistical process control). His scientific contribution with 99 publications (H=33) cited 3700 times is impressive for a non-academic. 

To proliferate and cultivate Chemometrics, scientists and engineers must be made aware of the benefits of the methods, have access to the methods and educated on the proper use. Barry has successfully achieved all three goals. Dr. Wise has through Eigenvector Research, Inc., using PLS_Toolbox/Solo taught thousands of students how to apply chemometrics, from basic methods like PCA and PLS to advanced non-linear methods like ANNs and SVMs. There are thousands of chemometrics users, from novices to experts, who have been impacted by the efforts of Barry Wise. These people span myriad industries from semiconductor, pharmaceutical, petrochemical, manufacturing, medical devices, agriculture, food and beverage and automotive areas. 

Barry has also received numerous awards and honorary recognitions including Chemometrics award at Eastern Analytical Symposium (EAS, 2001) and Chair of GRC on Statistics in Chemistry and Chemical Engineering (1995), and CAC-2002. 

Barry has dedicated 30 years to the chemometric community and is a true role model for students and scientists in sharing knowledge and through polite and logical argumentation in scientific discussion. 

In conclusion, we find that Barry Wise is a most worthy receiver of the Herman Wold gold medal of 2019. 

*Chemometrics is the chemical discipline that uses mathematical, statistical, and other methods employing formal logic to design or select optimal measurement procedures and experiments, and to provide maximum relevant chemical information by analyzing chemical data.

Hyperspectral Image Analysis Tools

May 30, 2019

When I was first introduced to image analysis by Paul Geladi he referred to it as Multivariate Image Analysis or MIA. So when we released our image analysis package back in 2005 we called it MIA_Toolbox. Since then more and more analytical techniques have been adapted to produce images. On the micro scale this includes surface analysis techniques such as SIMS and Raman microscopy. On the macro scale it includes remote sensing such as infrared imaging. Based on the expansion of the number of channels in the spectroscopic dimension it’s become more common to refer to the data as Hyperspectral images. Regardless of what you call it, MIA_Toolbox was built to handle it!

Early versions of MIA_Toolbox brought the conventional chemometric tools and preprocessing methods found in PLS_Toolbox to images, focusing the analysis more on the relationship between variables (typically wavelengths or mass) then on the relationship between pixels. In this it was a departure from conventional image analysis which focuses on the latter. For instance Principal Components Analysis (PCA) could be applied to images, where the neighboring pixels are treated independently, and then the results could be displayed back in the image plane as image score plots, such as the one below.

Of course many other methods could be applied in a similar fashion, including Multivariate Curve Resolution (MCR) and of course classification methods such as SIMCA and PLS-DA. It was also easy to do regression provided that reference values corresponding to the image were available.

Since then we’ve added techniques that add spacial information such as Maximal Autocorrelation Factors (MAF) which finds factors that are more highly correlated in the image plane. An example of MAF is show below and it can be seen that it creates images with more contiguous areas than the PCA image shown above.

Texture is a measure of the spacial variations of intensity in an image. This property can be important in the quality of some manufactured surfaces, in crystallization processes, or to assess the homogeneity of pharmaceutical tablets. Since 2009 MIA_Toolbox has included texture filters which can be used to create spectra from images that capture the spatial variation.

With the integration of the ImageJ image processing library into MIA_Toolbox in 2011 particle analysis was enabled. It is now possible to create a workflow where hyperspectral images of particles are treated with PCA to reduce dimensionality, then fed into particle analysis where each particle becomes a sample with measured size and shape characteristics, then those are pushed into a classification method to sort particles based on both their physical and chemical attributes. A screen shot from such a work flow is shown below.

In more recent releases we’ve continued to refine the useability of the tools, adding more file importers, preprocessing methods, etc. And of course all of this is also included in our stand-alone image processing package, Solo+MIA.

Want to learn more about image analysis? We’re doing courses at this year’s EigenU Europe in Montpellier, FRANCE, and the SCIX conference in Palm Springs, USA. At EVRI we’re excited to create tools that address the next generation of chemical analysis.

BMW

EigenU Poster Winners

Jun 22, 2017

Hello EigenFriends and EigenFans,

Two months ago we hosted the 12th Annual EigenU at the Washington Athletic Club in Seattle, WA. We had over 50 people join us through six days of hands-on chemometric courses, and once again had an awesome poster session where users got to display their use of PLS_Toolbox and EVRI methods in their recent research. We named two winners to take home the poster prizes from the session: Dr. Gordon G. Allison of Aberystwyth University, and Dr. Amanda Lines of PNNL. Information about their posters are attached below; congrats to you both!

21866_IBERS Seattle Conference Poster_FINAL1

Says Allison, “The Eigenvector PLS toolbox, and later the MIA toolbox, has been a staple of my data analysis platform for over 10 years. If it wasn’t the best I would have moved on. The software integrates seamlessly into Matlab and I move between command line and gui at will. The flexibility and variety of algorithms in the PLS toolbox never ceases to impress me, and Eigenvector seem to never run out of ideas of how the toolbox can be refined and expanded… I heartedly recommend that PLS toolbox users come along to the annual Eigen University in Seattle, to hone their skills in basic to advanced chemomentrics. I’ve been to 3 and always learn something new and valuable. The atmosphere is vibrant, fun, friendly and informal, the quality of instruction excellent, and the location couldn’t be better.”

Dr. Amanda Lines of PNNL also took home a poster prize regarding her work on “Using multivariate analysis to quantify and identify speciation of plutonium”. The abstract is below:

Amanda M. Lines1, Susan R. Adami1, Sergey I. Sinkov1, Amanda J. Casella1; Gregg J. Lumetta1, and Samuel A, Bryan1

1- Energy and Environment Directorate, Pacific Northwest National Laboratory, Richland, WA 99352

Development of more effective, reliable, and fast methods for monitoring process streams is a growing opportunity for analytical applications. Many fields can benefit from on-line monitoring, including the nuclear fuel cycle where improved methods for monitoring radioactive materials will facilitate maintenance of proper safeguards and ensure safe and efficient processing of materials. On-line process monitoring with a focus on optical spectroscopy can provide a fast, non-destructive method for monitoring chemical species. However, identification and quantification of species can be hindered by the complexity of the solutions if bands overlap or show condition-dependent spectral features. Plutonium (IV) is one example of a species which displays significant spectral variation with changing nitric acid concentration. Single variate analysis (i.e. Beer’s Law) is difficult to apply to the quantification of Pu(IV) unless the nitric acid concentration is known and separate calibration curves have been made for all possible acid strengths. Multivariate, or chemometric, analysis is an approach that allows for the accurate quantification of Pu(IV) without a priori knowledge of nitric acid concentration. Chemometric analysis is also an effective avenue for quantifying multiple species in solution that exhibit overlapping bands. This is demonstrated by the accurate measurement of multiple actinides (Pu(IV), Pu(III), U(IV), U(VI)) in multicomponent streams under dynamic conditions.

Thanks to everyone who participated in the poster session this year; all the posters were fantastic and it was dubbed the best EigenU poster session yet!

Top 10 reasons to come to EigenU 2017!

Mar 5, 2017

Howdy folks! All of us here at EVRI are getting excited for the 12th annual Eigenvector University, April 23-28 at the Washington Athletic Club in Seattle, WA. We’ve been updating our class list, and adding new examples and modeling techniques to make sure you get the best training in multivariate methods and chemometrics. Here are the top 10 reasons why we think you should be excited too:

  1. More experienced instructors – Courses at EigenU 2017 will be led by the EVRI staff including Neal B. Gallagher, R. Scott Koch, Robert T. Roginski, and Willem Windig, plus our Associate Rasmus Bro, and of course myself. Together we’ve got over 100 man-years of chemometric experience, and we’re all here to make sure you have a productive, fruitful week at EigenU.
  2. Wider variety of courses – In addition to our beginning track including PCA and PLS, we offer 11 advanced and specialty courses. This year features encore presentations of Managing Data, Models & Plots in PLS_Toolbox, and Model Deployment with Solo_Predictor & Model_Exporter. Calibration Model Maintenance returns this year with new and updated tools, as does the popular Bring Your Own Data (BYOD) Workshop where you work with your own data while you learn hands-on with EVRI’s team of instructors.
  3. Method-centric instruction – At EigenU we provide the background required to truly understand chemometric methods; we don’t just show you what buttons to push. Our goal is to make the literature in the field accessible to our graduates. Deeper understanding of the methods leads to better analysis!
  4. Spend a week in beautiful Seattle, WA – With Puget Sound and the Olympics to the West and Lake Washington and the Cascades to the East, the Emerald City is distractingly scenic. Plus, with attractions like the Space Needle, Pike Place Market, Seattle Art Museum, Seattle Mariners, and the largest ferry system in the US, there’s plenty to see and explore while not in class.
  5. The Washington Athletic Club – EigenU is held at the WAC, the nation’s premier city athletic club. The historic 21-story facility includes 5 floors of fitness facilities, 10 floors of Euro-styled techno-centric sleeping rooms, full service spa, and 3 restaurants.
  6. The food – From the continental breakfast, including the WAC’s signature sticky buns, the gourmet plated lunches, to afternoon snack bars, our guests always rave about the food.
  7. Networking – EigenU attendance is typically about 40 scientists and engineers with a range of chemometric expertise and wide variety of interests. This means you’ll have plenty of opportunity to find colleagues with common problems and complementary solutions.
  8. Evening events – EigenU provides ample opportunity to continue your chemometric learning and networking into the evening. This includes Tuesday’s PLS_Toolbox/Solo User Poster Session, Wednesday’s PowerUser Tips & Tricks Session, and Thursday’s Workshop Dinner, which is one more opportunity to enjoy the WAC’s fabulous food. The best two posters on Tuesday evening’s User Session will win a pair of Bose wireless headphones or wireless speakers!
  9. Flexible, multi-platform software – With PLS_Toolbox and MIA_Toolbox EVRI offers the most comprehensive set of chemometric tools available plus the flexibility of MATLAB. Our stand-alone packages Solo and Solo+MIA offer all the point and click tools of their MATLAB-based siblings. Plus they’re all available for Windows, Linux and MacOS. On-line tools Solo_Predictor and Model_Exporter provide a plethora of options for automating model application.
  10. Costs less – In spite of all its advantages, EigenU actually costs less than similar courses from our competitors, making it a reasonable and affordable option for everyone interested in honing their chemometric and advanced data analysis skills!

So it probably isn’t surprising that EigenU attendees are more than satisfied. Here’s what a couple of them had to say:

“EigenU is a wonderful, positive and informative place to learn about chemometrics methods that I feel confident applying to not only my research, but suggesting to collaborators as well.” -Brooke Reaser, University of Washington PhD Candidate.

“What you are offering here is unmatched.” – David A. Russell, Dupont.

Spots are already filling quickly so register today to reserve your place at EigenU 2017! Early registration ends March 23.

See you in April!

BMW

EigenU 2016 Poster Winners

Oct 6, 2016

Each spring at Eigenvector University in Seattle we showcase the scientific accomplishments of our users at the PLS_Toolbox/Solo Poster Session. With EigenU Europe quickly approaching, we thought it was about time to congratulate our poster winners from EigenU 2016 last April!

Brooke Reaser of the University of Washington, and Claire Muro of the University at Albany took home the poster prizes at the 11th Annual Eigenvector University. Brooke’s work used principal component analysis to determine the uptake of 13-C-labeling in Methylobacterium extorquens, while Claire’s poster focused on the differentiation of body fluid traces using PLSDA models. Their posters are attached below; congrats, ladies!

Forensic Body Fluid Differentiation Poster
claires-poster“PLS_Toolbox doesn’t just help with my research, it is an integral part of all of my work. I work with Raman spectroscopy, and my experimental data for any given project is usually thousands of spectra. I would say that PLS_Toolbox is just as important to my research as my Raman spectrometer, and I mean that. I wouldn’t be able to process the data or learn anything from it any way other than multivariate data analysis.” -Claire Muro, University at Albany PhD Candidate.

PCA C-13 Labeled Methylobacterium Extorquens AM1 poster
brookes-poster-screen-shotEigenU helped give me the idea to apply PCA to my metabolomics project in the first place, and the skills I learned there helped me in applying it successfully and meaningfully. EigenU is a wonderful, positive and informative place to learn about chemometrics methods that I feel confident applying to not only my research, but suggesting to collaborators as well.” -Brooke Reaser, University of Washington PhD Candidate.

For their efforts both Claire and Brooke took home Bose sound system products. Claire selected the Bose QuietComfort Headphones while Brooke took home the Bose Soundtouch Speaker System. We hope you enjoy them, you earned them! Congratulations again!

STAY CONNECTED WITH US!
Eigenvector now has an Instagram account! Follow @eigen_guys to keep up with the latest adventures of the Eigenvector team, and make sure to check out our Twitter account for up-to-date info on courses and software.

Register for EigenU Europe, October 24-27

Oct 5, 2016

Registration for EigenU Europe in Montpellier, FRANCE is open! Join Dr. Barry M. Wise, Prof. Rasmus Bro and Dr. Sébastien Preys for four days, October 24-27, of beginning and intermediate chemometrics courses from the classic Eigenvector University series. The course will include:

We’ve also added a follow-on event Friday, October 28 with Dr. Fabien Chauchard on How to Implement Spectroscopic Techniques for Process Development.

43020496-c01a-467d-929b-dba8a2612bf9

To register and view lodging information, visit the EigenU Europe page. We hope to see you there in beautiful Montpellier next month!

STAY CONNECTED WITH US!
Eigenvector now has an Instagram account! Follow @eigen_guys to keep up with the latest adventures of the Eigenvector team, and make sure to check out our Twitter account for up-to-date info on courses and software.

What Works and What Doesn’t

Jun 28, 2009

After SSC11 I got a note from my colleague Paman Gujral at the Automatic Control Laboratory at EPFL summarizing some of the talks. He wrote: “Rasmus Bro gave an excellent talk too about the pitfalls in using chemometric methods. Kowalski commented that software firms are a lot to blame for advocating methods that don’t work.” I was a little alarmed by this and so asked Rasmus about it, who wrote: “The comment of Bruce is maybe correct but it wasn’t meant as he states it. … as part of a discussion Bruce and others mentioned that also software companies have responsibilities in helping people take proper decisions. This was added to a more general agreement that education should be improved. So there was nothing dramatic or controversial in that.

In any case I’ve thought a great deal about that comment since Paman’s note. In fact, as a software company sometimes it’s hard to know when a new method comes out if it is going to live up to it’s initial hype. This is largely because nobody publishes negative results. If you lived on a steady diet of J. Chemo and ChemoLab, you’d think that EVERYTHING works, at least initially. So yes, it’s often up to us to sort out what works and what doesn’t. The often means coding it up ourselves, and trying it out. And sometimes even then the results are ambiguous. We don’t see where it definitely doesn’t work, and feel that, given what was in the original journal article, it’s worth putting into the software.

We figure that if more people have experience with a new method then eventually we’ll all figure out if it is generally useful. That happens faster when there is code available which implements the method. So yes, I’ll admit that we’ve put things in PLS_Toolbox that were “unproven.” I have stated this publicly, more than once. That said, I don’t recall ever promising these new techniques would work. Caveat emptor. Only in this case, it’s “let the user beware.”

I’ve put a lot of effort trying to make some methods live up to the initial claims. I could give you a list of my all time greatest “wastes of time” but at this point that would only serve to upset the originators of the methods. But I think we’ve been of more than a little service in helping sort some of these things out.

I certainly agree that “education should be improved,” and we strive to do that. One of the things we tell our students is to not believe everything they read or hear, and we try to give them the tools to dig past the hype. We also teach proper model validation. If you do a proper validation, you’ll at least know if one of these new techniques doesn’t work on your particular data.

But I don’t know how many times I’ve had to answer the question, “Why doesn’t your software do ______? The author/speaker says that it’s the best thing since Gauss.” And I have to answer, “Well, we haven’t tried it. But our experience suggests there are no silver bullets.”

BMW

Update on Chemometrics in Rome

Aug 6, 2008

Details are now available for our Basic Chemometrics in Rome, Italy, short course. Rasmus Bro will be joining me for this 3 day, hands-on course October 27-29, 2008.

The course page includes additional information about Rome, the course content, registration, prices, and nearby lodging. If there is anything you can’t find, please let me know!

BMW

Sabbatical at EPFL

Jul 16, 2008

In September of 2007 I received an offer I couldn’t refuse. When Professor Dominique Bonvin of Ecole Polytechnique Fédérale de Lausanne (EPFL) wrote and asked if there was a chance to have me spend some time there in 2008, it didn’t take long for me to send back a positive response.

Besides the fact that EPFL is in a beautiful location, (on Lake Geneva in Switzerland), Dominique’s group there at the Automatic Control Laboratory has research interests that are well aligned with Eigenvector’s, and, of course, our users. For instance, they have worked on issues surrounding MSPC and curve resolution for investigating reacting systems.

I’ll be working with Michael Amrhein on model updating schemes for batch processes. This is rather timely as we work more and more with monitoring batch processes, including semiconductor manufacturing and bioreactors. Model updating is important because these process exhibit considerable drift with time.

I’ll be spending a little more than 2 months at EPFL, arriving in the last couple days of August and leaving in the first few days of November. My wife (and assistant!) Jill, will be along, as will daughters Clare and Mattie. It should be pretty much business as usual at EVRI, though it make take a few days on either end to set up operations to process orders, etc.

I look forward to this time to “sharpen the saw.”

BMW