NFL Draft Assist
Persistent URL
Author(s)
Joseph II, Vital
Date Issued
May 6, 2025
Abstract
NFL Draft Assist is a decision support tool designed to assist in NFL draft process decisions and analysis. It uses predictive modeling to evaluate college players entering the National Football League (NFL) draft and predict which ones will succeed at the next level. That evaluation process incorporates a variety of metrics like, college production, physical body measurements, combine statistics and statistical techniques like linear models, monte carlo cross validation score thresholding, and clustering. What was found is that those metrics don’t just add up in a straightforward way to determine a player’s potential. It takes a combination of statistical analysis, manual feature engineering and a deep understanding of what really correlates to success in the NFL to get a accurate read on a player’s chances of success. That’s where the real value of this project lies. The model that performed the best, Linear Threshold Weighted, achieved an 80% accuracy rate in classifying current NFL players by how successful they were likely to be before the NFL draft and compared that to how well they are doing in the NFL right now. This accuracy shows that a data-driven approach might be useful to the scouting process. What it also reveals are the problems in the way teams currently draft players. You see biases toward athleticism over actual performance, and the way certain positions and schools get overvalued or undervalued. That’s where this tool comes in to play. NFL Draft Assist is a web application built with Flask that lets users visualize players predicted draft outcomes for upcoming NFL draft, as well visualizations to see how well the models perform on current NFL players. This tool is a resource for scouts, media professionals and analysts to make more informed decisions and talking points regarding NFL draft prospects. By integrating predictive analytics, like NFL Draft Assist, into the traditional scouting process, teams can understand potential bias and make informed draft decisions. This is cutting edge in an industry where data is increasingly driving decisions about NFL talent.
Major
Computer Science
First Reader(s)
Graber, Emily
Other Reader(s)
Jumadinova, Janyl A.
Department
Computer and Information Science
Type of Publication
Senior Project Paper
File(s)![Thumbnail Image]()
Name
SeniorThesis.pdf
Size
871.57 KB
Format
Adobe PDF
Checksum (MD5)
9ebf59ce345f79593a83bd34ed87a920