March Edition: Data and Causality

Author:Murphy | View: 25088 | Time: 2025-03-23 19:34:08

Monthly Edition

In a recent Author Spotlight Q&A, Matteo Courthoud reflected on the growing importance of making robust predictions, whether one works in industry or in academia:

I think in the future, causal inference will become more and more central and we will see a convergence between the theoretical approach from the social sciences and the data-driven approach from computer science.

We hope you read the rest of our lively conversation; in the meantime, Matteo's observation inspired us to dive into our archives in search of other insightful articles on Causal Inference and the topic of causality more broadly. The resulting selection we're sharing in this Monthly Edition goes from the introductory to the more advanced, and showcases some of the different approaches data science and ML practitioners use every day in their work.

We hope you enjoy exploring these recommended reads! As always, we're grateful that you've made TDS part of your learning journey; if you'd like to support our work in other ways (and gain access to our entire archive along the way), please consider becoming Medium members.

TDS Editors

TDS Editors Highlights

The Science and Art of Causality (Part 1) (January 2023, 11 minutes) Quentin Gallea, PhD‘s accessible introduction to the fundamentals of causal inference aims to answer two key questions: why is understanding Causality so important, and why are causal relations so hard to assess?
Causal Inference Using Synthetic Control (November 2022, 9 minutes) In situations where A/B testing isn't a good option for establishing causality, quasi-experimental techniques might be the answer. diksha tiwari proposes synthetic control as one such alternative.
Causal Effects via Regression (January 2023, 8 minutes) Shawhin Talebi has been covering the theory and practice of causal inference for over a year. You can go all the way back to the beginning of the series, or jump straight to this recent post on regression techniques and how we can use them to establish the relationships between variables.
Event Studies for Causal Inference: The Dos and Don'ts (December 2022, 17 minutes) Event studies are another helpful approach in quasi-experimental contexts; as Nazlı Alagöz points out in this well-explained deep dive, there are many pitfalls to avoid so that we don't draw the wrong insights from our data.
Statistical Significance Testing of Two Independent Sample Means with SciPy (November 2022, 8 minutes) For readers who are keen to roll up their sleeves and start experimenting with some data, Zolzaya Luvsandorj‘s tutorial is an ideal starting point: it provides a practical entryway into hypothesis testing, and includes all the Python code you'll need.
Identification: The Key to Credible Causal Inference (February 2023, 8 minutes) As Murat Unal explains in an incisive new post, "without a well-defined identification, no amount of sophisticated modeling or estimation can help us in establishing causality from data." Read Murat's overview to gain a better understanding of identification and its importance.

Original Features

Explore our latest selection of Q&As and reading recommendations.

"The main driver behind my writing has always been learning"Our Q&A with Matteo Courthoud, in which he reflects on leaving academia, his interest in causal inference, and the value of public writing.
Confronting Bias in Data Is (Still) Difficult – and NecessaryWe shared essential readings on a topic that continues to dominate conversations within ML and AI communities.
How to Build Good Habits as a Data ScientistIn a recent roundup, we compiled tips and strategies from seasoned data practitioners for more streamlined and reliable workflows.