A Multivariate Statistical Analysis of the NBA

Will your favorite National Basketball Association (NBA) team make it to the playoffs this year? What variables affect a team’s postseason outcome? In an attempt to determine which teams will make the NBA playoffs, we will collect and analyze team data using multivariate statistical methods including Principal Components Analysis and Discriminant Analysis. Introduction It is midway through the NBA season. The owner of the Milwaukee Bucks takes a look at the standings; it is a close race in the Central Division. The one question weighing on his mind, will we make the playoffs this year? With so much revenue to be made by clubs as well as enjoyment to be had by the fans, making the playoffs is vital to any team. Is there a way to determine with relative certainty whether or not a given team will make it to the playoffs? Division standings are likely to change throughout the season, so are there other components that one can use to predict which teams will make it? The goal of this paper is to analyze data from the NBA in order to gain insight into the questions asked above. We begin by constructing a list of variables that we think are important for any team to make the playoffs. Then, we use Principal Components Analysis (PCA) on these variables to find new components to describe our data. A key result of PCA is that we reduce the dimensionality of the data set. We also use Discriminant Analysis on our variables to predict the classifications of teams into one of two populations, playoffs or non-playoffs. Data We chose 12 variables that fall under the categories of game statistics, player demographics, team finances, coaching, and fan support. The data comes from various websites containing information on the NBA. Some variables are taken as is and others needed further calculations. The following is a table of variables and how they are calculated: