Level: Intermediate
Akira Hososhima, Representative Director, Torin Consulting
Previous examples of using JMP to solve quality problems have included reliability prediction methods, defect cause analysis methods, and quality trend analysis using text mining. This time, we'll introduce a method for using the analysis of mean percentages (ANOM) to extract and narrow down model numbers and equipment with significantly higher defect rates from production performance data, and an example of modeling operator setup know-how in a virtual experiment using a deterministic screening design. When analyzing data that is added regularly, such as production data, preparing fixed data such as product model number master data in a separate table and linking the two facilitates updates. We also introduce an example of an initiative that reduced errors by one-tenth annually by visualizing human error. We graphed the total number of operational errors and the breakdown of the error content, and used graphs stratified by individual, position, and organization to provide feedback to supervisors and operators to encourage awareness and innovation. We'll also introduce how to use Graph Builder to drop stratification variables into Group X, columns, and overlays, as well as stratification using Local Data Filters (LDFs).
After serving as Director, Head of Research and Development, and Head of Quality Assurance Promotion at Yamatake-Honeywell (now Azbil), he founded Torin Consulting. His areas of expertise include quality improvement, reliability prediction, robust design, general statistical problem solving such as experimental planning, design review, root cause analysis (RCA), and human error prevention.
His major publications include "Practical Success Strategies for Startup Companies" (Chuokeizaisha, 2011) and "Easy-to-Understand Book of Problem Solving" (Nikkan Kogyo Shimbun, 2014) (sole author). His major papers include "Proposal for a Workshop to Promote the Dissemination and Acceptance of Awareness of Near Misses and Unusual Events on Production Lines" (2016 Society of Quality Control Quality Technology Award winner). His major lectures include "Proposal for a System to Visualize Organizational Factors that Trigger Work Errors and Promote Improvement" (Discovery-Japan, 2018), "Solving Quality Problems with JMP: Case Studies of Defect Analysis and Reliability Prediction" (same publisher, 2019), "Improving the Efficiency of Experiments and Analysis with JMP" (same publisher, 2020), "Method of Investigating Quality Problems on High-Mix, Low-Volume Production Lines Using Text Mining" (same publisher, 2021), and "Reforming Development, Production, and Quality Assurance Operations with JMP" Parts 1-5 (JMP On Air Japan Edition, 2021).
