This has been a busy year for Pega Analytics. I’d like to take a few moments to review what we have been doing, and what is in the pipeline.
Over the summer we built a course that was specifically designed to align with the book Visual Six Sigma. First delivery of the course has just taken place to rave reviews! The course was designed to be modular, so we’ll be looking to deliver individual components of the course, for example, running a module on Definitive Screening Designs.
In addition to delivering training courses Pega Analytics specialises in the JMP Scripting Language, and we’ve been involved in some really interesting projects during the year. One such project has been the development of an Ambr Connector for importing data from a TAP Biosystem into JMP.
The new website was built from the ground-up using an entirely new technology. It is based on what is known as a “responsive design”. The idea is that the content and layout adapts to the size of your browser. That’s important, particularly as your browser may be on a tablet or mobile phone.
The site incorporates the use of some interesting “widgets” that facilitate user interaction. My favourite is the course selector widget that delivers a really effective way of browsing courses when you are on a device such as an iPad.
New Introductory Course for JMP
For many years Pega Analytics has run a two day course: Getting Started With JMP. Whilst it has been very successful, there was clearly an interest in a similar course but running over a single day. How do you compress two days content into one day? The trick was to recognise that this was really quite a mature course. In fact, it was the first course that we produced. It has had iterative updates to reflect changes in JMP. So we took the themes of the course and rewrote the content of the course from scratch.
“I think the different ways of learnings (quiz, exercises , videos, demos) kept the audience engaged”
Visual Six Sigma
This course gives us the opportunity to teach some interesting topics ranging from Wheeler’s EMP Method to Definitive Screening Designs. All the topics are discussed within the context of 5 case studies that are used to illustrate real-world scenarios from beginning to end. The common thread that underpins the course is the workflow shown above and as outlined in the book. Traditional approaches to teaching six sigma rely on complex workflows that are designed to validate the assumptions that underpin statistical analysis. This approach reinforces the concept of “statistics as a lawyer”. Most statistics courses use this formulaic approach to teaching. It feels like a safe but lazy approach to teaching a complex topic. Statistics is a component in a broader theme of data analysis. And the act of doing data analysis is a process. And if it is a process then we can seek to make that process lean. That’s what visual analytics is all about. It’s what JMP is all about. So Visual Six Sigma is about using the JMP-way of doing things, as applied to the data analysis problems that we see when we conduct six sigma (or lean) projects.
Ambr250 is a fully automated bioreactor system for microbial fermentation and mammalian cell cultures. Pre-sterilised disposable bioreactor vessels can be controlled by an automated workstation to enable execution of experiment runs. Automated execution and data collection is great, but what do you do if you want to analyse the data in JMP?
Let’s say that you want to create an overlay graph of oxygen rate and carbon dioxide evolution for each of the 24 vessels used to conduct an experiment. The data for each variable is stored in a separate file. And distinct files are maintained for each bioreactor vessel. So that means to create the overlay graph you will need to import data from 48 files. And the files are not even stored in the same location; data for each vessel is stored in a separate folder. Once you have imported the data you will have 48 separate JMP files, but you only want one. The 24 files for each variable need to be combined and then data for the two variables needs to be joined. Each data point has a timestamp which needs to be converted to an age to convey useful information. And if you pay close attention to the data you will discover that some of the most recent data values are missing. Ambr stores “current values” in a separate file structure. You could try reading this into JMP but it won’t be easy because it is a binary file. And all of this assumes that you can find the data files in the first place. But that is not always easy. They often have cryptic names like this: “4163dc85-d82a-40b0-82d9-da86934d1e94 64748659CA6542F3116D85CE4F74140FF3E21300.all.csv”. So first a variables identifiers file needs to be referenced to determine the contents of each file.
The Ambr Connector solves the problem. You select the variables that you are interested in and click the import button. Job done. Now the data is in JMP as a single table. You can now use the full power of JMP to investigate the data.
We’ll be looking to make a refresh to our website. The refresh will include dates for SAS Public Training Courses in 2017. The Visual Six Sigma course will be added to our course catalog and we’ll look to develop some other courses based on components of the Visual Six Sigma course. In particular:
- Measurement Systems Analysis
- Modelling Workflow
- Definitive Screening Designs
We do a lot of work on JSL and the website only gives you a superficial glimpse of that work. We’ll be looking to provide more detail of the projects that we are undertaking as well as the free addiins that we make available on the JMP file exchange.