Code Yarns β€πŸ‘¨β€πŸ’»
Tech Blog ❖ Personal Blog

The Art of Statistics

πŸ“… 2021-Oct-29 ⬩ ✍️ Ashwin Nanjappa ⬩ 🏷️ book, statistics ⬩ πŸ“š Archive

I regularly look for book recommendations from folks I admire. When both Terence and Deepak talked about The Art of Statistics I decided to read it straightaway. The author David Spiegelhalter has spent his entire life at the Royal Statistical Society in UK, aiding government, courts, hospitals and corporations with his weapons of statistics. And classic British humor is everywhere in this book that reads like a whodunit or problem solver, motivating all fields of statistics with real and famous examples from history or the author’s past.

The early chapters (1 and 2) introduce summary statistics like mean, median and standard deviation. Graphs are also introduced and all manner of graphs remain a constant feature throughout the book. Chapter 3 features random sampling and normal distribution, which is used in Gallup polls. The following chapter gets into the complexities of how randomized controlled trials are conducted for new drugs. How do we distinguish between corelation and causation - that is discussed a lot with real examples. In chapter 5, we see regression models and how everything regresses to the mean.

We get to the current hotness of machine learning in chapter 6 with a cursory look at classification trees, ROC curves and such. The author stresses on the challenges of using algorithms due to problems like bias and hard to interpret. We are far from summary statistics and in the land of probability theory in chapter 8. And in the following chapter, statistics and probability are used together for statistical inference. The latter chapters get into statistical significance and P-values and Bayesian probability. The book ends with a broad look at statistics are misused and guidelines for how to do better.

I felt the author made the book a lot of fun to read with his humor, showing the history behind every statistical innovation and famous statistician and most importantly the great real-world examples used to motivate every concept. The biggest takeaway from the book for me were those examples, witnessing the messiness and complexity of applying stats in the real world in medicine, criminal cases, courts and government policies. The book also has a good glossary with formal definitions of all the statistical terms and a detailed bibliography from which I added several more books. The book achieves the breadth, helps you learn about all the different fields which come under the wide umbrella of statistics. This is not a book that can be used to study and learn them, though the author tries to make it easy - you’d need a real statistics textbook to do that. This was a thoroughly entertaining and enlightening read and has motivated me to read more math-flavored books in the future.

Rating: 4/4