Statistical Thinking for Industrial Problem Solving (EU 2018 413)
Feb 16, 2018 9:19 AM
| Last Modified: May 7, 2018 10:36 AM
Data Analysis Assistant.jmpaddin
Level: Intermediate Mia Stephens, JMP Academic Ambassador, SAS Martin Demel, JMP Systems Engineer, SAS Ian Cox, JMP Senior Marketing Manager, SAS
In response to feedback and requests from both academics and industry leaders, JMP is developing an online course in statistical thinking for industrial problem-solving. In this breakout session, we introduce the principle of statistical thinking and present a practical, data-driven approach to solving real-world problems that serves as the foundation for this new course. We show how to exploit the visual and interactive nature of JMP for exploratory and confirmatory data analysis, and explore core techniques and modern statistical methods through examples and simulations. An end-to-end case study is used to unify concepts and broaden applicability, and a preview of the online course will be provided.
A note on the materials:
If you are running JMP 14 download the Zip file for JMP 14. This contains the all of the materials packaged in a JMP Project. This will unzip in a jmpprarchive file, which will create a Project (.jmpprj file) and a folder structure C/DEMO/Discovery/2018_StatisticalThinking/ which includes all of the content.
If you are running JMP 13 download the Zip file for JMP 13. This includes the journal, with separate scripts for the analyses, and the pdf for the talk.
Update March 19:
During the talk we discussed the development of a "Data Analysis Assistant". This will help users determine the statistical method(s) that make the most sense for a given analysis task. This is designed for the non-statistician - engineers and analysts who may have limited statistical background. So, the analysis task is in non-technical language, "what do you want to do?"
A prototype add-in and the data table that feeds the add-in are attached. Please note that this is a very early version - the data table is not comprehensive, and the add-in design is an initial prototype. However, we'd welcome feedback and suggestions on both the concept and the best way to provide this sort of tool.