In this project we seek to conduct a statistical study of the NBA, from its beginnings in 1940 to the present, we seek to show the evolution and how the trends of the game, teams and players have been adapting over time.
The main objective of this study is to analyze the league in general, from the competitiveness of the different seasons and establish correlations between the individual awards given to players and their statistical performance, such as performing a principal component analysis or PCA to reduce the number of variables and finally a predictive model for the MVP.
The first thing is to examine the evolution of the players over the years. We will analyze variables such as shooting percentage from the field, rebounds and assists. To do so, we have used statistical analysis techniques to identify trends and patterns in the data.
Also, we have studied the evolution of the teams over time. Here we analyzed variables such as number of wins, average points scored, and points received. And we have also identified dominant teams in different eras.
One of the most interesting aspects of this study is the evaluation of the competitiveness of each NBA season. For this purpose, we have used competitiveness indexes, such as the HICB or the sigma index (see reference [7]). Basically, we compared the seasons with each other and determined which ones have been competitive.
In addition, we perform a principal component analysis to reduce the number of principal components that determine the performance of players and teams.
Added to all this, is the predictive model of the MVP we have done it using Python, more specifically with the Pandas library to obtain the DataSet needed to train the model, and sklearn another library where basically it is used to make the model and obtain the prediction.
On the other hand, we have examined the correlations between individual awards given to players, such as MVP (Most Valuable Player), Defensive Player of the Year, SMOY (Best Sixth Man), and their statistical performance. The use of statistical techniques to determine which variables have a greater relationship with obtaining these awards has helped us in obtaining and analyzing this data.
Finally, we have identified which players within the top performers of the different statistics have been more determinant for the team's victories.