I’m always on the look-out for quality resources that I can recommend to my beginning data science and machine learning students. Last year, I wrote a review of a useful book that got students up to speed with a key mathematical ingredient of machine learning – linear algebra: The Manga Guide to Linear Algebra. No Starch Press (an excellent source of technical books) just came out with a follow-up title: The Manga Guide to Regression Analysis by Shin Takahashi and Iroha Inoue. Using Japanese manga (comics) as a framework, the book provides a delightful introduction to necessary topics that many newbie data scientists might find difficult such as logarithms, Euler’s number, differential Calculus (a cute way of calculating a derivative is included), matrices, etc. In addition, the book covers some basic topics in statistics such as hypothesis testing, variance, standard deviation, and probability density functions.
If you’re light on a math background, I think reading manga is an entertaining way to fill in your knowledge. Most of the important books that data scientist need to consume require familiarity with the math behind linear regression, arguably the most popular supervised machine learning technique. Chapter 2 covers simple regression (one predictor) and Chapter 2 discusses multiple regression (multiple predictors).
Reading this book might be a nice prelude to diving into a statistical programming environment like R, since topics like ANOVA, confidence intervals, residuals, R-squared, multicollinearity etc. will make a lot more sense. As an added bonus, the book covers binomial logistic regression which is another popular supervised learning method designed to predict probabilities whether or not something will happen.
The book is rather brief with only 216 pages and four chapters:
Chapter 1: A Refreshing Glass of Math
Chapter 2: Regression Analysis
Chapter 3: Multiple Linear Regression Analysis
Chapter 4: Logistic Regression Analysis
I already have added a slide to the regular presentation deck I use at conferences for my “Data Science Primer” talk, recommending this book for folks just starting out in the field. I must give kudos to No Starch Press for thinking up this innovative way of teaching a potentially difficult subject. After all, comics have a disarming effect even if the subject is math!
Contributed by Daniel D. Gutierrez, Managing Editor of insideBIGDATA. In addition to being a big data journalist, Daniel is also a practicing data scientist, educator and sits on a number of advisory boards for various start-up companies.
Sign up for the free insideBIGDATA newsletter.