Causal Inference and Discovery in Python: Unlock the secrets of modern causal machine learning with DoWhy, EconML, PyTorch and more Paperback – 31 May 2023

I highly recommend this to any data professional in digital marketing or clinical research. It clarifies foundations in a tight and approachable way to non-causal data scientists. I found it quite practical with priority on intuition before formality when breaking down theory.

The media could not be loaded. I love it🌟

I'm currently about 2/3rds of the way through this book and I've found it to be an excellent read. The author makes the complex simple with his well referenced and researched writing. Absolutely recommended for anyone interested in getting up to speed. It covers the basics as well as the most recent techniques like Double Machine learning, causal forests and more. Interested in causal analysis, then don't hesitate. But it!

Was a present.

This book is a game-changer for anyone interested in making sense of causal machine learning (ML). You don't need to be a causal inference pro to get a lot from this book, but a bit of Python and ML background helps.Here's why I loved it:It explains complex stuff in simple ways. Even tricky topics like how things cause other things and imagining "what if" scenarios are made easy to understand.It's packed with real-life examples using Python tools like DoWhy, EconML, and PyTorch, showing how to apply what you learn.The writing is clear, making tough topics easy to get.The book is perfect for folks who want to understand the why behind data, not just the what. It's especially cool for those looking to explain their AI work better or integrate it into fields like healthcare where knowing the cause is important.Highly recommend it if you're into data science and want to explore beyond the usual machine learning techniques!

*I've read every line of a decent chunk (~60%) of this book now - this review is based on Sections 1 and 2.Positives:There value in this book - especially if you've got a grounding in the theory of causal inference/discovery, but want to learn more about practical solutions and packages in the Python ecosystem. There are comprehensive references (including very recent arXiv papers). The light touch on theory does keep you moving through at a rapid pace. The book itself is decent enough quality.Negatives:There are quite a few places where things could be tightened up. The book (in it's present form) has a very unusual style: something like slides from a intensive course on causality annotated with a script and references. While this conveys the gist well enough, a lot of time is lost reading poetic motivations, while basic technical details are omitted, or referenced (sometimes as youtube videos). For me personally there is too much fluff and unnecessary verbalisation of equations, the space of which could have easily been repurposed for explanation / fundamental technical content. There are some loose ends: the description of Simpson's Paradox in Chapter 1 is left without any closure as a mystery - while one of causal modelling's big wins is that it can resolve paradoxes like this.Overall:I think I would buy this book again - the references, survey of algorithms, and code, provide decent enough value. BUT, for me, it's still a distance from 4 star in the present form.