2019 draft machine learning predictions

This repository contains the code and methods for each post analyzing the 2019 draft. Each post has its own folder with its own data, code, and graphs.

6/20: Predicting the best scorers in the 2019 draft with machine learning

The first project is predicting the best scorers in the 2019 draft. It uses rookie and college data since the 1990 draft and projects PPG for prospects who played in college and are projected in the first round of the Ringer's mock draft (as of 6/15 update). It can be accessed in the "scorers" folder.

Link to blog post.

7/8: Generating stats-based historical comparisons for the draft lottery

The second project uses cosine similarity and Euclidean distance to create comparisons for the lottery. It uses college data since the 1990 draft and computes similarity between each player picked in this year's lottery and the entire historical database of lottery picks. It can be accessed in the "similarity" folder.

Link to blog post.

7/26: Using machine learning to predict All-Stars from the 2019 draft

The third project uses four classification models to predict All-Stars from the top-10 picks of the 2019 draft. It uses college data from the 1990-2015 drafts to predict All-Stars. It can be accessed in the "all-stars" folder.

Link to blog post.