Seeqing Automated Model Retrains with Data Lab – INVISTA

Stephanie Ma, Modeling & Analytics Engineer, INVISTA

In this breakout session, Invista share how they used Seeq Datalab and AI-assisted Python to automate model retraining for rotating equipment at Invista’s Houston plant. They explain how her team reduced nuisance alerts by automating engine changeout checks, adding stabilization buffers, filtering downtime and outliers, and generating cleaner model bounds for more reliable failure prediction. The session also explores future enhancements—including prediction intervals, SHAP values, and automated model selection—and offers practical guidance for non-computer scientists on using Datalab and AI to build repeatable modeling workflows that better serve asset owners.