Practical Machine Learning in R

First the good part: this is an overall excellent book. The conceptual explanations are very clear and understandable, at beginner level. The book covers all the usual algorithms. regression, classification, unsupervised learning. There are early chapters on R and data wrangling. The book doesn't actually use a ton of packages, just the tidyverse and a handful of specialized libraries. The code blocks and their outputs are well explained.
So why not 5 stars?
So, I basically re-typed all the code from the book and I often ended up with different results especially for logistic regression and classifications. Sometimes it's to be expected but I used the exact same seed so, I should have gotten the same results. Part of the issue is that the dataset (which are provided with the book, you just have to download them from the website) are not exactly the same as they are in the book. Some had different numbers of rows.
The other bigger problem was that, in the classification problems, we often had to deal with class imbalances. To address that, the book uses a library that no longer exist (Torgo's DMwR). There is a DMwR2 library that goes with the 2nd edition of DMwR but it does not include the SMOTE function that is used throughout the book.
Now, it's not an impossible problem. There are other ways of addressing class imbalance. However, choosing a different method then guarantees different results. Mine were closed except for logistic regression.
That being said, if you teach that stuff to beginners, the book is worth it because of the clarity of the explanations. You can always fix the rest. But don't just assign the book without the proper corrections.