# Route Identification in the National Football League

## TL;DR

This paper presents a model-based clustering approach using Bernstein polynomials and EM algorithm to identify receiver routes in NFL tracking data, enabling better analysis and querying of game plays.

## Contribution

It introduces a novel method for classifying NFL player trajectories into routes using curve clustering with Bernstein basis functions and EM, improving play analysis.

## Key findings

- Successfully clustered 34,698 routes from 6,963 plays.
- Enabled new receiver metrics considering deployment.
- Facilitated streamlined game film queries.

## Abstract

Tracking data in the NFL is a sequence of spatial-temporal measurements that vary in length depending on the duration of the play. In this paper, we demonstrate how model-based curve clustering of observed player trajectories can be used to identify the routes run by eligible receivers on offensive passing plays. We use a Bernstein polynomial basis function to represent cluster centers, and the Expectation Maximization algorithm to learn the route labels for each of the 34,698 routes run on the 6,963 passing plays in the data set. We go on to suggest ideas for new potential receiver metrics that account for receiver deployment. The resulting route labels can also be paired with film to enable streamlined queries of game film.