Independent Data Science Project using NBA shot data
Using R, I analyzed data from 10,000 NBA shots. The goal of the project was to model and predict field goal percentage. I discovered trends to increase field goal percentage by up to 20%. Using the Caret package, I performed high-level feature and model selection to create a Naive Bayes model that predicted shot precision with 65% accuracy. I created various visualizations to effectively display how field goal percentage varied. I also used the package Shiny to build an interactive web tool to help coaches easily manipulate variables to see how effective shots were in different shooting zones. The tool can be accessed at: https://jasonk33.shinyapps.io/NBA_Field_Goals/
The entire write-up can be viewed here: https://drive.google.com/file/d/0B274UXOwQJrSZmxqZnpkN01MUjg/view?usp=sharing