The book is aimed at those who truly want to professionally engage with Python applications in data science and all its subsets (Machine Learning, Deep Learning, etc). It is partly good that it does not use the libraries scipy, scikit-learn, and tensorflow. Instead, it implements simple statistical metrics and relatively complex classifier models (Naive Bayes) in detail. This is the best approach for someone who has no prior experience in data analysis, as it allows them to see the basic mathematical tools that play a role. The book assumes basic knowledge of Python, and anyone who understands, understands. It comments on each method and occasionally provides advice. It might overwhelm you a bit with the frequent use of functions, but in the end, it will prove useful as you see how to write reusable code. Additionally, you will never be asked to write knn, k-means, or naive bayes from scratch like Grus does. Once you finish this book, you will be able to start exploring the scipy, scikit-learn, and tensorflow libraries and call each model with just one command. You may ask why then should you read this book that does everything from scratch? I would say that a data scientist should be a good mathematician-statistician, in addition to being a competent programmer. Therefore, by understanding the mathematics behind classification, clustering, and regression algorithms, you will gain an understanding of each algorithm's concept, which method is best suited for specific problems, and of course, how to optimize each method. Personally, I would have liked more comments and advice targeted at working with data (preprocessing, filling empty values) and dense clustering algorithms. In conclusion, if you partly agree with what I'm saying, take it and you don't need to read all the chapters right away. I recommend chapters 1 to 18 and 20.