Exercises to learn JMP

[#1]

View more...

In this thread I will collect and add some exercises I use in JMP Days or Workshops to have an easy to grab location for people joining those sessions.

This is a living thread I update from time to time with new/adjusted material or new information. There are many good resources to learn JMP (like STIPS, Welcome Kit, Mastering JMP, Trainings, ...) and this thread is no replacement for these. It really is to have one location I can point to in my sessions where I have some homework/hands-on activities.

Alternative Material:

A collection of Learning Resources for JMP Users

Case Study Library 

The interactive Resource Path

Essentials Using JMP Session

We use the "Tablet Production.jmp" Sample Data Set in JMP (Help -> Sample Data ->See alphabetic List of all Sample Data Files

• Session Exercise: 
  • How much data is outside the specification limit of the response "Dissolution"? Use e.g. Control Chart Builder
  • Using Graph Builder, what are the potential drivers on Dissolution? Are there some potential nonlinear effects, in what variables?
• Homework Exercise:
  • Proof that Screen Size has significant impact on Dissolution! Use e.g.  Analyze -> Fit Y by X. Then recreate Control Chart for "Dissolution" and filter by Screen Size. Note the differences in the out of spec results for Dissolution.
  • What would be a setting to avoid running out of spec? Use e.g. Analyze->Predictive Modeling->Partition platform to understand the combined effects driving less out of spec settings

Essentials for Designing an Experiment

This is the use case:

• Session Exercise: 
  • Create a screening design with the custom designer leaving the default settings
  • Analyze the results using the "Model" Script as well as Graph Builder to understand what factors seem to have an impact on yield. (using "Chapter 2 results.jmp" attached in this post)
• Homework Exercise:
  • Create the same screening design in the custom designer as in the exercise, but with minimum runs as well as with more runs. What can you say about the Power of the design? What design would you choose when you have to decide?
  • Read through What is correlation? and What is regression? in the statistical knowledge portal to understand what the p-value represent and what a linear model is.
  • In preparation of the next session about Screening and Definitve Screening Designs read through this blog post  by Bradley Jones, one of the inventors (next to Chris Nachtheim) of Definitive Screening Designs.

Essentials about Screening and Definitive Screening Designs

We discuss this use case:

These are the factors considered:

• Session Exercise: 
  • Create a Definitive Screening Design using the attached factors table ("DSD Factors.jmp") and the default settings
  • Analyze the results in the attached JMP data table "DSD Results.jmp" using the "Fit Definitive Screening" Script
  • Create different traditional Screening Designs with about the same number of runs and compare these to the DSD approach
• Homework Exercise: 
  • Review proper use of Definitve Screening Designs by reading through this blog post  by Bradley Jones, one of the inventors (next to Chris Nachtheim) of Definitive Screening Designs.

RSM, Robustness, and QbD

Files: Factors.jmp - Robustness DOE Example.jmp - Optimal Robust Example.jmp

• Session Exercise: 
  • https://www.jmp.com/en_ca/events/ondemand/technically-speaking/analytical-method-development.html 
• Homework Exercise:
  • https://community.jmp.com/t5/JMP-Blog/The-QbD-Column-Applying-QbD-to-make-analytic-methods-robust/ba...

DOE with Categorical Factors

Homework:

• Watch Isn-t-Discrete-Numeric-just-Categorical 
• Read The-other-5th-thing-I-wish-someone-had-told-me-when-I-started