Our new episode with Robert J. Reynolds is out now! 1/ #CausalSky #StatSky #EpiSky #MLSky

Last week I went to New York to talk to Robert J. Reynolds, PhD, for a new podcast series with @sobor.bsky.social and UBC. Rob had quite a story to tell about his work with causal models at NASA. #CausalSky #StatSky #MLSky

We just sent today's issue of Causal Python Bi-Weekly. Here's what's inside: 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

I'm in NYC to record a couple of new episodes of the #CausalBanditsPodcast What questions should I ask Elias Bareinboim and Robert J. Reynolds? 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

AI assistants can convince us to make bad decisions while we keep believing they were helpful. And it's a significant problem in statistics, causal analysis, and research. 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

Peeking at A/B tests? Does the stopping rule actually matter? In today's issue of Causal Python Weekly: 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

"AGI has arrived" (But not for the first time) #CausalSky #AISky #MLSky #StatSky

I'm a huge nerd when it comes to nutrition. #CausalSky #StatSky #EpiSky #NutriSky #MLSky

In a few days, Yale University Press releases Scott Cunningham's Causal Inference: the Remix. But you can already read it for free online. 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

The idea behind difference-in-differences is ingenious. But... Let's start from the beginning. #CausalSky #StatSky #EconSky #EpiSky #MLSky

Continuous treatment difference-in-differences with debiased machine learning in Python? Why not?

When you rebuild a randomized trial from observational data, how often do you actually get the same answer? In this week's issue of Causal Python Weekly: 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

Negative controls are an everyday staple in epidemiology. This framework based on them is a secret in industry. 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

Evidence-Based Decision-Making & Causality What do you want to learn next? Over the last few months, I've been asking people in our community what would be the most useful resources for them when it comes to their learning goals. 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

Interesting!

Connor T. Jerzak@jerzakconnor.bsky.social · 2mo ago

Planetary Causal Inference combines satellite data with local studies like RCTs. This fall at UT Austin, join GOV 385L: Causal ML & Earth Observation for Social Science to explore the field. Tuesdays, 8:30–11:30am CT.

In today's issue of Causal Python Weekly: 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

Time to reveal our special guest! Today, we spent a great day on the Massachusetts Institute of Technology campus with Justin Bélair Tomorrow, at the Thomas Menino Center, we'll be joined by Professor Mark van der Laan 1/ #StatSky #CausalSKy #MLSky #EpiSky #EconSky

In Boston? Join us tomorrow for a session on causality and complex dynamic systems! Complex systems have a peculiar property: their emergent behaviors can act back on the lower-level entities that produce them, while also impacting emergent properties on arbitrary levels. 1/ #CausalSky #StatSky

I always felt that Python has a significant gap when it comes to experimental work. Sure, SciPy handles simple designs, power, or sample size for a t-test. But more complex designs? Those almost always required custom code. 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

A Tool Is Just a Tool. And Isaac Gerber took it to another level. On Monday, we're releasing the final episode of Season 2 of #CausalBanditsPodcast. And it's going to be quite a finale! 1/ #CausalSky #StatSky #EpiSky #EconSky #MLSky

Can an LLM stand in for the humans in your A/B test? Mårten Schultzberg, Joel Persson, and Sebastian Ankargren take the question seriously and work out what it would actually require.

How much value is your experimentation program leaving on the table? This week in Causal Python Weekly: 1/ #CausalSky #StatSky #EconSky #EpiSky #MLSky

How PC-Based Methods Err (and How to Detect it) Causal discovery methods built on the PC algorithm are sound, but only if every structural assumption holds and every conditional independence test is correct. 1/ #CausalSky #EpiSky #StatSky #MLSky #EconSky

What Does It Take to Trust a Causal Discovery Result? In this week's issue of Causal Python Weekly: 1/ #CausalSky #StatSky #EpiSky #EconSky #MLSky

In these personally challenging times, it feels nice to know that what we've been doing for the last 2+ years with the podcast brings value to the community. I just got a message from YouTube that Causal Bandits Podcast hit 100k views. Thank you for your continuing support! ❤️ #CausalSky #StatSky

Our standard predictive machinery requires statistical repetition. Can we say anything about the effects of one-time or first-time events? Alexander's model correctly predicted the behavior of oil prices in response to the events in the Strait of Hormuz 1/ #CausalSky #EconSky #MLSky #StatSky

"History doesn't repeat itself, but it rhymes." The new episode of the Causal Bandits Podcast with Alexander Denev is out now The recent closure and reopening of the Strait of Hormuz are unique events. 1/ #CausalSky #EconSky #EpiSky #MLSky #StatSky #FinTech

Predicting the Effects of One-Time Events, a DML Case Study, and More... Tomorrow, we premiere a brand-new conversation with Alexander Denev, meanwhile... ...in today's issue of Causal Python Weekly: 1/ #CausalSky #StatSky #MLSky #EconSKy #EpiSKy

Is Causal Inference Dead in the Age of Vibe Coding? This week at Allegro Tech Talks, I'll discuss 5 causal ideas that I believe we should be aware of in the age of vibes. We'll discuss: 1/ #CausalSky #StatSky #EpiSky #EconSky #MLSky

Causal sensitivity analysis is expensive at scale. Could a foundation model fix that? In this week's issue of Causal Python Weekly: 1/ #CausalSky #EconSky #EpiSky #StatSky #MLSky