Multivariate Statistical Analysis on NBA Finals Team

Background

image.png

-Data is always changing
-Change of the rule
-A phenomenal player causes a butterfly effect
-An outstanding team
-Show what changes on NBA finals teams data have experienced

Objective

image.png

Method-Data Filtering

image.png

Method-Multiblock PCA models

image.png

Results & Discussion-Multiblock PCA

image.png

image.png

image.png

image.png

image.png

Results & Discussion-Multiblock PCA(4X)

image.png

image.png

Conclusions

image.png

-1980 Los Angeles Lakers may be the best team in NBA finals through the data.
-Three point has become a trend in NBA.
-We cannot say a team playing at home and doing well in defensive rebounds would win the championship, however, if they play away and have terrible defensive rebounds, they can hardly win.

Deficiencies & Future Work

-Filtering and analyzing data is largely artificial. Bias is inevitable and it is -possible to draw false conclusion.
-There are many flaws in our data in terms of the contribution of indicator to the game.
-We believe there are more than simply digital statistics in today's sports. It must have record data with vectors. Such as a shooting distribution on the court, different frequency of tactics, or even distribution of the pass between players. All the questions are expected sports fans like us to figure out.