Best Sellers Rank: #88,823 in Books (See Top 100 in Books) #5 in Computer Programming Logic #53 in Python Programming #172 in Artificial Intelligence & Semantics
Customer Reviews: 4.4 out of 5 stars 114Reviews
Product Information
From the Publisher
Why is causal inference such a key topic for data scientists to learn about?
In 2022 there were an average of 3.2 new papers on causality published on ArXiv every day, a number which has been growing exponentially over the past 3-5 years. Top researchers and organizations like Microsoft, Amazon, and DeepMind invest their resources in causal research and we are seeing more and more causal applications in industry. Companies across various business sectors implement causal methods – from gaming to manufacturing, from finance to automotive - and among them are companies like Spotify, Playtika and BMW.
This book will help you learn about causal inference by covering the basicsnecessary to understand this new and dynamic field, and – using a step-by-step approach – we then move towards more advanced and state-of-the-art methods, helping you to build a comprehensive, and powerful skillset.
Table of Contents
Causality – Hey, We Have Machine Learning, So Why Even Bother?
Judea Pearl and the Ladder of Causation
Regression, Observations, and Interventions
Graphical Models
Forks, Chains, and Immoralities
Nodes, Edges, and Statistical (In)dependence
The Four-Step Process of Causal Inference
...and more!
What was your objective in writing this book?
When I was starting my journey with practical causality, I could not find a comprehensive book on causality in Python.
Understanding the potential of causal machine learning and knowing how much effort it took me to build my skill set, I wanted to share my journey with others, so they can enter this dynamically evolving field easier and faster and start applying causal inference and causal discovery in their own projects.
What is your favorite part of the book and why?
I enjoyed working on all parts of the book, but I have a special fondness for chapters 7 and 11. The former introduces the idea of the 4-step process of causal inference. This is an idea that originates from the DoWhy package created by Amit Sharma and colleagues, and I believe it’s one of the most powerful ideas to help newcomers build a clear structure around the causal inference process.
In chapter 11, we discuss the intersection of causality and natural language processing (NLP), which lays the foundation for understanding fascinating recent research on causality and generative AI. My bet is that we’ll see dynamic growth in this area in the coming years, and so this chapter can prepare the reader to more easily grasp the new ideas in the field and apply them quickly.
What are the key takeaways from this book for readers?
I see three main key takeaways for the readers. The first is general in its nature and it’s about causal thinking. Causal thinking is thinking in terms of the data-generating processes rather than statistical summaries of the data. I see it as one of the most powerful data skills in the upcoming 3 to 5 years and I am confident that it can help virtually anyone become a better data scientist, analyst or researcher.
The second takeaway is that working with causal models doesn’t have to be scary or exceedingly difficult. It boils down to a set of practical and mental skills that can be learned by anyone, and my hope is that the book does a good job in helping you achieve this. The last takeaway is that by giving ourselves a space for creativity, we can face and overcome even the most difficult challenges. I see practical causality as a beautiful example of this phenomenon.