2025-10-01 –, Louis Armand 2 - Ouest
Causal inference offers a principled way to estimate the effects of interventions—a critical need in industrial settings where decisions directly impact costs and performance. This talk presents a case study from Saint-Gobain, in collaboration with Inria, where we applied causal inference methods to production and quality data to reduce raw material usage without compromising product quality.
We’ll walk through each step of a causal analysis: building a causal graph in collaboration with domain experts, identifying confounders, working with continuous treatments, and using open-source tools such as DoWhy, EconML, and DAGitty. The talk is aimed at data scientists with basic ML experience, looking to apply causal thinking to real-world, non-academic problems.
This talk describes the implementation of a causal inference framework at Saint-Gobain Research, in the context of optimizing a glasswool production process. The goal was to identify which operational factors causally influence a target variable (thermal conductivity delta), and ultimately to reduce raw material consumption while preserving output quality.
The talk is intended for data scientists with experience in Python and ML, but without prior exposure to causal inference. We will provide an end-to-end walkthrough of a causal study on observational data, covering:
• Collaboration with domain experts to build the causal graph
• Challenges of identifying confounders from available data
• Working with continuous treatments (vs. binary interventions)
• Estimation strategies using DoWhy and EconML
• Visualizing dose-response curves and exploring heterogeneous effects
• Applying refutation tests to assess robustness
• Communicating findings back to domain experts through Conditional Average Treatment Effects (CATEs)
We’ll reflect on practical difficulties (e.g. expert collaboration, graph validation), share lessons learned, and emphasize reproducibility using only open-source tools.
I am data science project leader at Saint-Gobain Research in Paris. From 2018 to 2022 I worked at the Aramis team (INRIA) where I obtained a PhD in computer science from Sorbonne University. My PhD was about machine learning for 3D neuroimaging using a large scale datawarehouse. I obtained a bachelor and a master degree in biomedical engineer from Politecnico di Torino
