Basketball Data Science

With Applications in R

Paola Zuccolotto, Marica Manisera

• Publisher: Chapman and Hall/CRC
• Year: 2020
• eBook ISBN: 9780429894251
• Edition: 1
• Topic: Economics
• Subtopic: Statistics for Business & Economics
• Index: Economics

Using data from one season of NBA games, Basketball Data Science: With Applications in R is the perfect book for anyone interested in learning and applying data analytics in basketball. Whether assessing the spatial performance of an NBA player's shots or doing an analysis of the impact of high pressure game situations on the probability of scoring, this book discusses a variety of case studies and hands-on examples using a custom R package. The codes are supplied so readers can reproduce the analyses themselves or create their own. Assuming a basic statistical knowledge, Basketball Data Science with R is suitable for students, technicians, coaches, data analysts and applied researchers.

Features:

• One of the first books to provide statistical and data mining methods for the growing field of analytics in basketball
• Presents tools for modelling graphs and figures to visualize the data
• Includes real world case studies and examples, such as estimations of scoring probability using the Golden State Warriors as a test case
• Provides the source code and data so readers can do their own analyses on NBA teams and players

I

Getting Started Analyzing Basketball Data

Chapter 1

Introduction

ACCORDING TO Marc Gasol, statistics are killing basketball. The Spanish big man gave his opinion about statistics almost at the end of the regular season of the NBA 2016/2017, when, according to statistics, he had just become the first center to record 300 assists, 100 threes and 100 blocks in a single season.

We’ve got 43 wins. If we win (tonight), we’ll have 44. That’s the only number you guys (media) should care about. Stats are great, but wins and losses matter. Stats are killing the game of basketball. Basketball is a subjective game. A lot of things happen that you cannot measure in stats. Different things matter. To me, the most important things in basketball are not measured by stats.

Marc Gasol
#33 Memphis Grizzlies 2016/2017

Let’s start our book on basketball Data Science by providing a disconcerting revelation: according to us, Marc Gasol is right. What we are talking about is Statistics and Data Science in basketball and how to make it a useful tool but not, as Marc Gasol fears, as a way to reduce the game to numbers that are not truly able to describe it. So, let’s start from the wrong way of understanding Sports Analytics. First, the vast majority of people believe that Statistics in basketball can be reduced to counting the number of shots, baskets, points, assists, turnovers, …. And, in a way, this simplification makes sense: every day the specialized media report news about these so-called statistics, detected in the NBA games, and fans are delighted to bet on whom will be the first player who will exceed this or that record. But these statistics (and we deliberately continue to write the term with a lowercase initial letter, to distinguish it from Statistics, which is the science we are dealing with) don’t say much, and Gasol is right. For a Statistician, these are simply data that, collected in large quantities and appropriately re-elaborated, can be transformed into useful information to support technical experts in their decisions. Surely, to evaluate a performance only on the basis of these values is not only very reductive but even, in some cases, misleading. In addition, we think that the concept that Marc Gasol tried to convey is one of the most crucial: the media, with their often sensationalist claims, spread the statistics as if they were the thermometer of the players’ skills and success. In this way, players will sooner or later change their way of playing with the goal of keeping high stats, with all due respect to teamwork. And the only real goal is to win. Actually, the opinion of the Spanish player is not limited to this topic, already fundamental in itself, but adds another aspect that is even more subtle: are the numbers really able to describe the game? How could we measure aspects such as the way a point guard is able to control the pace of the game and his decision-making skills, the influence of a leader on the team’s self-confidence, the cohesion of the players, the extent to which defense is firm and tough, etc.? Certainly this is not counting the number of assists, points, and steals. You do not need to be an expert to understand it. On this point, unexpectedly, Statistics (the one with a capital S) has a lot to say because there is a well-developed research line that deals with the study of latent variables, that is all those variables that are not concretely and physically measurable. The tools invoked by these methods are sophisticated techniques and algorithms. So, if Big Marc ever read this book, he would find out that he is right: the most important facts of basketball are not measured by statistics. But they could be measured by Statistics and, in general, by Data Science.

1.1 What is Data Science?

Data Science is the discipline aimed at extracting knowledge from data in various forms, either structured or unstructured, in small or big amounts. It can be applied in a wide range of fields, from medical sciences to finance, from logistics to marketing. By its very nature, Data Science is multidisciplinary: it combines Statistics, Mathematics, Computer Science and operates in the domains of multivariate data analysis, data visualization, artificial intelligence, machine learning, data mining, and parallel computing. In fact, several skills and abilities are required for a Data Scientist: he needs ...

Table of contents

• Cover
• Half Title
• Series Page
• Title Page
• Copyright Page
• Dedication
• Contents
• Foreword
• Preface
• Authors
• Part I: Getting Started Analyzing Basketball Data
• Part II: Advanced Methods
• Part III: Computational Insights
• Bibliography
• Index