Ahead of the Curve, Analysis of the NFL Draft: Technical Report

The National Football League (NFL) draft is the primary gateway to the professional ranks of American football. Any college football player worth their salt dreams of being picked number one in the first round of the NFL draft. This project attempts to predict the likelihood of whether a player entering the draft will be selected in the first round. Previous statistics from the National College Athletic Association (NCAA), NFL combine results and previous draft results will be analysed following the Knowledge Discovery and Data Mining methodology. Data will be targeted, selected, transformed and modelled to reveal knowledgeable patterns which will aid in the prediction. Who, apart from avid fans, cares about this? Well, the NFL is a multi-billion-dollar industry made up of 32 teams and the individuals who choose team membership are the General Managers (GMs) of these teams. These GMs are faced with the decision at the start of each season of who to pick for their team; should they retain the players they have or choose the hot new prospect? Techniques such as data visualization and principal component analysis were used to explore the datasets. Regression was used to develop models for the prediction, packages such as Glmnet and XGBoost provided smooth model development. The data was then visualized in a user-friendly dashboard using Tableau