Topic: A Statistical Engineering Example: Forecasting Stability at P&G

Speaker: Fangyi Luo, Director/Principal Statistician and Data Scientist, P&G

Date: Tuesday Jan 27, 2026 at 8:30 am PST /11:30 am EST / 5:30pm Central European Time

Abstract

In this webinar, I will share a statistical engineering example from Procter & Gamble, focusing on the forecasting of product stability and predicting shelf life or long-term stability failure risk with uncertainty estimates at the early stages of product development. This was an unstructured, large, and complex problem and we used a statistical engineering approach to tackle this problem effectively.

I will also provide an overview of various statistical methods used in forecasting and predictive stability, including degradation and reliability modeling, chemometric Multivariate Curve Resolution, Bayesian Analysis of Differential Equations, AI deep learning, Bayesian Network modeling, and Functional Data Analysis.  I will demonstrate applications of these methods across multiple products and stability failures.  Additionally, I will share key learnings gained from solving this complex problem, as well as some of the newer challenges we are currently working on.

Speaker Bio

Fangyi Luo, Director/Principal Statistician and Data Scientist, has been with P&G for 28 years. She holds a Ph.D. in biostatistics from the University of Cincinnati and has statistical design and modeling experience across various areas, including clinical and consumer research, formulation, process, stability, and packaging. She has extensive expertise in developing advanced predictive and forecasting stability models and is currently focused on packaging modeling and 3D computer vision.