Moving from Dashboards to Decisions: Using Modeling to Improve Your Metrics (2026-US-30MP-2918)

In manufacturing, more operational data is being collected than ever before. Metrics are being calculated for availability, performance, quality, etc., and are broken down by days, product line, and shift. All of this is widely available in dashboards, often throughout the manufacturing floor. Production managers are able to slice the data so they can see exactly how and why production performed good/bad yesterday, last week, last month. But when asked how to improve their metrics, they point to the biggest bar in a Pareto chart or give vague answers about reducing errors and doing better. This can result in chasing a lagging indicator, an erroneous problem, or a one-time issue – all of which means wasted resources.

A better approach is to apply predictive models and forecasting to see what will happen in the future and create an asset management plan for improvement. In this presentation, we showcase the tools in JMP that can be used to analyze operational data and techniques for forecasting with this data to identify exactly where to focus efforts for improvement. This presentation covers:

• How to use JMP tools to clean operational data so it's ready for analysis.
• Different types of models to consider for prediction.
• Forecasting with the data to estimate the future; identify the impact of improvements.

The presentation concludes with a demonstration of how this process was applied to real operational data and the insights that were gathered.