Behavioral Data Analysis with R and Python: Customer-Driven Data for Real Business Results

Why do you analyze data? Why do you run a regression? Why do you build a model?Is it to make better predictions under the assumption that the future will be the same as the past, or is it to understand the causes of current phenomena (whether the data source is experiments or observations) and answer what-if (business) questions? If your intellectual curiosity pushes your mind more toward the second question, this is the book that you will thank yourself for deciding to read.On top of that, do you want to see how to apply a causal-behavioral framework to a business question? Sometimes, don't you wish there was a way not to overlay that famous normal distribution data assumption onto your work but to use your data as it is? What about learning a solution when your testing unit and test executable unit are different? And wouldn't it be cool to get a clear roadmap to locate experiments (tests) in the journey of causal inference combined with a behavioral analytics approach? The list of useful information and ideas you can get from this book can go on and on, and these are just a few examples.You will find that the author did an excellent job of presenting not-easily understandable concepts in a reader-friendly way. Additionally, you will also see that the author shares a lot of useful practical tips that must have come only from real working experiences and kindly answers the very questions readers may already have in the book. Again, this book will surely satisfy your hunger for intellectual amusement and provide business applicable knowledge on the spot. So grab it and enjoy it!

easy to comprehend and very useful for real-life data analysis!

This book is a fantastic resource for those in the field who are interested in causation and robust data analysis. Speaking as a Sr. Data Scientist, I found nearly every page of this book both approachable and relevant. I've recommended it to virtually everyone on my team and in some sense I think it should be required reading for the role.Regarding the delivery of the content, Buisson strikes a healthy balance between rigor and accessibility that I think will cater to a large audience in the field. There is some math but it's not a textbook and the intuition is always clear. Regarding the content itself I almost cannot speak highly enough. In terms of causal inference, many books cover DAGs, chains, colliders, the backdoor criterion etc. but few provide a sensible framework for how to actually go about constructing these DAGs in a given context. Buisson does just this and the end result feels highly tractable (in lieu of some gigantic DAG where there is a cause to everything). Other chapters that I found highly relevant covered bootstrapping, all things experiments, moderation (interaction) and mediation (instrumental variables). All of which I have already found helpful in my actual day to day as data scientist.There are many topics this book does not cover, so do not go in thinking this is a one stop shop. But perhaps the best feature of this book is the content that Buisson does cover that appears absent in so many others. Buisson is an applied scientist and this experience shines through brightly. If you are in the field and you're not just building black box models (cool if you are but you'll know if this is not you) then you might just love this book.Books to pair this with: Anything by Angrist and Pischke. Something that covers the Rubin Causal Model / potential outcomesPrereqs: Some stats and probability. Familiarity with linear regression. Basic R and Python.

Quoting Eric Weber (Yelp Head of Experimentation):"This book is unique in that it starts with the questions and problems and leverages the techniques and programming languages as true tools. Readers will learn to solve incredibly important and tricky problems. It is well worth your time and investment."The only thing I can add is that it is an easy and fun read since it is written in a conversational style. This is a technical book for professionals in the field, but it is written in such a way that people who do not consider themselves experts can still benefit from it. I highly recommend this book for anyone interested in behavioral data analysis.

The book is well written. Florent makes complex concepts accessible to readers with clear explanations and code.

The author takes something not difficult, word vomit on the pages, and at the end, (1) you understand less (2) legit irritated.Explanation and writing style are convoluted.Who’s the target audience for this book???

Behavioral Data Analysis with R and Python: Customer-Driven Data for Real Business Results is an excellent resource for anyone interested in behavioral data analysis. The author, Florent Buisson, strikes a perfect balance between education and accessibility, making the book suitable for a wider audience than most similar texts. As someone working in research and experimentation, but with no background in behavioral analysis nor R or Python - I had a lot to learn and was a little worried I'd be overwhelmed quickly. Instead, I was able to follow along, even smiling at a few of the analogies thrown in along the way to help keep me entertained.Well worth the purchase and the read. Highly recommended.

The field of "behavioral data science" is increasingly popular and exciting. It's also ill-defined. Florent's book provides a deep and detailed look at how behavioral science and data science can be effectively united: using techniques from simple regressions, to causal diagrams, to missingness models, he shows us how to understand the specific causes of a behavior. It's good. It's nuanced. It's thoughtful. If you want to understand the data footprints of human behavior, this is the book you need.

Temas y programas muy básicos