Causal Inference in Practice: Beyond Classical A/B Testing

🙂 This is an in-person event 🙂
⚠️ (Ideally) Register on LinkedIn: [https://www.linkedin.com/events/7429445270809112576/]

We all love a clean A/B test, but let’s be honest: the real world is rarely that cooperative. Whether it’s interference between users, ethical constraints on randomization, or non-compliance, there are times when classical experimentation simply breaks down.
How do leading data teams bridge the gap? They turn to other causal inference methods.
In our 6th session, we’ve invited top practitioners to pull back the curtain on how they identify, estimate, and validate causal effects when a standard randomized controlled trial isn’t an option.

💬 What We’ll Cover

• The "Why" Over the "What": Concrete use cases where A/B testing failed and Causal Inference saved the day.
• The Methodology Deep-Dive: From mapping causal structures with Directed Acyclic Graphs to estimating effects via Difference-in-Differences, how do you choose the right method?
• The Tech Stack: A look at the go-to tools, libraries, and custom code (Python/R) used to build these models.
• Operationalizing Insights: How to scale causal analysis and communicate uncertainty to stakeholders who are used to "simple" A/B test results.

🎤 The Format
This session is a hybrid of deep-dive technical presentations and a candid panel discussion. No slide-heavy lectures here—just real-world examples, robustness checks, and honest talk about what happens when the data is messy.

- Speaker: **Xiaowei Zhang — Manager Data Science (Causal Inference Consulting & Research) @Booking.com**
- Speaker: Antanas Mainelis — Lead Decision Data Scientist @Vinted
- Moderator: Jos Baan — Senior Data Scientist @Spotify

✨ Why Attend? ✅ Master the transition from experimental to observational causal analysis. ✅ Learn how to handle "interference" and "non-compliance" in your data. ✅ Network with a community of peers who are navigating the same technical hurdles. ✅ Get a look at the diagnostic checks that determine if an effect is actually "identifiable."

🍕 Practicalities
- Drinks & Food 🍷 🍕
- In-person only: Because the best insights happen in the hallway and over a slice of pizza.
- Interactive: We’ve reserved plenty of time for Audience Q&A—bring your toughest causal questions!

⚡ Limited Seats To keep the discussion high-quality and interactive, we cap attendance. Make sure to RSVP early to secure your spot!

We can’t wait to see you there and continue building the ADU community together! 🚀

#AmsterdamDataUnion #CausalInference #DataScience #MachineLearning #ABTesting #DataEngineering #AmsterdamTech