The StatQuest Illustrated Guide To Machine Learning

I recently picked up a copy of Joshua's new book "The StatQuest Illustrated Guide to Machine Learning" (SIGML).And I must say, I'm very impressed 🤯!By a large margin, it is the most accessible book on ML that I've encountered...the anthesis of a typical dry, esoteric, & unintuitive ML book.Many technical and academic books alienate a large portion of their readers, self-sabotaging their educational value to those would-be learners by employing an esoteric vocabulary that's only accessible to people who possess specific academic backgrounds.By contrast, rather than making assumptions, Joshua takes pains to equip readers with a working knowledge of the language required for the concepts that he introduces - an approach that the entire technical and academic publishing sphere could learn a great deal from (i.e., focus less on sounding smart and more on helping people of all backgrounds to learn effectively)!Along this vein, I was surprised to learn something new in the very first chapter 😊!Many years ago, when first learning spreadsheets, I was introduced to data in rows and columns.Then, I moved on to structured databases with "tabular" data where the rows were referred to as "records" and the columns were referred to as "attributes".Later, in the ML space, I once again encountered tabular data...and this time the rows were called "observations" and the columns were called "variables" (which can be "independent" if they're informing the prediction or "dependent" if they're what's being predicted).By the time that I got into learning Stats/ML, I was mostly just amused to find yet another set of nomenclature for tabular data. I don't recall ever reading or being told why the fields of Stats & ML refer to the columns as variables (the discussion always focused on the independent vs. dependent part). So, without thinking about it much, I just accepted that in the ML context columns are called variables.Yet, here Joshua thoughtfully takes time to explain that columns of data are referred to as variables because the data "varies" from one observation to the next 💡.While completely logical, I have to admit being ignorant about this rationale for the nomenclature until yesterday when skimming through Chapter 1 of SIGML.This is a great book for those who're looking for a gentle, fully accessible introduction to ML that doesn't cut corners...it's also a good resource for seasoned ML practitioners who might want to go back and inspect their knowledge base for unrealized blind spots from a new, more illustrated perspective 😉.

For anyone seeking a intermediate\beginner-friendly and visually engaging introduction to machine learning, Josh Starmer's book is an excellent option. It effectively simplifies complex concepts, presenting them in an easily digestible format, often enhanced by clear illustrations. By minimizing the use of heavy mathematical jargon, the book creates a welcoming environment for readers who may feel intimidated by the subject matter. This approachable style makes it a valuable resource for newcomers to the field.

Best book to learn machine learning from basics ro complex algorithms. Josh have teaches the most complex algorithms in very simple way. Even a beginner with no knowledge can study it.

Josh has mastered one of the hardest skills in writing (and teaching): Explaining advanced concepts in an easy-to-understand way without leaving out important technical details. He does not dumb the algorithms and concepts down, but rather explains them in a more visual and down to earth manner, all while keeping the book fun and interesting to read. Squatch and Norm are great additions to the book as well, making it feel like you're on a learning journey together with them.Overall an amazing book. Josh is a far better teacher than my professors, and so not only has the book been good entertainment, but also a great help in machine learning courses at my university.

Illustrations are great and explanations are outstanding and of the same quality as the videos on the youtube channel

Really good content easy to understand.

The corners of the book came slightly warn.

The author has a clear understanding of what steps are needed to systematically teach ML. The illustrations and the arrows navigate the user through them while teaching concepts and building a strong foundation. Kids can understand these concepts more naturally since they study similar topics daily. But more importantly, the adults who have no domain expertise can also benefit immensely from this style of teaching. I am going to recommend this book to anyone interested in learning about ML