Possible Replacements for Philadelphia Eagles’ Wide Receivers Using Similarity Assessment
Introduction:
The Philadelphia Eagles have one of the best wide receiver rooms in the entire NFL. With the help of perfectly timed trades and great draft selections, the Eagles have the potential of having this receiver unit being a threat to opposing defenses for years to come. The current top 3 receivers on the Philadelphia Eagles are A.J. Brown, DeVonta Smith, and DeVante Parker. A.J. Brown was traded to the Philadelphia Eagles from the Tennessee Titans during the 2022 NFL draft. Having already been established as one the best wide receivers in the NFL, this was a blockbuster trade and a great new addition for the Eagles. DeVonta Smith was the 2020 Heisman winner. This award is given to the best player in college football that year. In the 2021 NFL draft, the Philadelphia Eagles selected DeVonta Smith with the 10th overall pick. WR3 for the Eagles is DeVante Parker. He was previously WR2 on the New England Patriots last year. His experience and athletic abilities brings a lot of football knowledge into their wide receiver room.
The player management operations of the Philadelphia Eagles is performed by General Manager Howie Roseman. He was responsible for all these trades, draft selections, and free agent signings. A large portion of his job revolves around the financials of the team and players. Paying these players based off their worth and competitive offers is key in having a great team. As players improve and have better performing seasons, their worth increases and therefore the amount of money they should be earning. As receivers are not the only positions to be considered, there are chances the wide receiver leaves the team because they simply can not afford to pay them the money they deserve. This is what I plan to investigate in this report. The motivating question is who the Philadelphia Eagles could potentially replace the 3 wide receivers they currently have with but still have the best shot at keeping the same level of performance. The similar wide receivers based on the similarity assessment would ideally have a lower salary than the existing 3 Eagles receivers. Using this dataset and analysis, this can help any NFL team, but I am choosing to focus on the Philadelphia Eagles. The main stakeholder of this report is Howie Roseman, the General Manager for the Philadelphia Eagles. The answers to the similarity assessment can help Howie Roseman make actionable decisions during free agency to trade and acquire players.
Data and Method:
The ideal data that I used for answering this question is a dataset from FantasyPros. This was a CSV file with the 2023 statistics of all NFL wide receivers. The features that were used in this analysis were receiving yards, targets, receptions, touchdowns, yards per reception, and 20+ yard receptions. These statistics are adequate representations of receiver output, explosiveness, and play style Previously, I analyzed the correlation between NFL receiving yards and touchdowns scored for all 2023 NFL receivers. I found that although there was a strong correlation, it was not as strong as many would expect. There were times where receivers had a lot of receiving yards, but not as many touchdowns and vice versa. This is probably due to the way the teams utilize the receivers in play schemes, but does not take away from the players’ performance potential. To focus on players that had some sort of performance output in the 2023 season, I took out the null values of the dataset because this dataset also considered practice squad players that do not have in game statistics. In addition, I cleaned up the data types of the columns to align with the similarity assessment. The receiving yards column (YDS) had a dtype of object- an example of a value in this column is 1,769. However, I changed this dtype to float and took out the comma, making this example value 1769.
Aforementioned, I built off the Module 1 assignment which was a surface-level exploratory analysis of receiving yards and touchdowns. This analysis performed basic data loading, cleaning, manipulation methods and insightful visualizations. To build off this, I incorporated a cosine similarity assessment for the 3 Philadelphia Eagles wide receivers.
Analysis:
Cosine similarity was the best similarity assessment for this case because it does not take into account the distance between the vectors like Euclidean similarity. Rather, it takes into account the angle between the vectors. In terms of football wide receivers statistics, this will not take into account unavoidable factors, like mid-season injuries, a lack of playing time due to personnel issues, or poor quarterback performance. If a player’s statistics are lowered overall due to these reasons, they can still be similar to one of the best receivers in terms of performance output when given a chance to play.
The most similar players to A.J. Brown and DeVonta Smith are the same players, in different order. This makes sense in football because A.J Brown and DeVonta Smith were both healthy, high performing receivers with a top 5 quarterback in Jalen Hurts. In terms of insights for decision making for Howie Roseman, a few names on this list are intriguing. One is Puka Nacua. He was a rookie wide receiver for the Los Angeles Rams last season. He was not regarded as a valuable asset in the draft, having been selected 177 overall. The Rams also had the previous NFL best receiver in Cooper Kupp. However, last season he had an explosive year, becoming the number one target on the Los Angeles Rams and one of the best in the NFL. Since he is a rookie, he is on a rookie contract which is substantially less than a veteran player’s contract. Howie Roseman can work his magic in trying to trade for him to replace either A.J. Brown or DeVonta Smith and pay him much less than he already does for these two. His output would generally be the same because the quarterbacks for the Eagles and Rams are both top 10.
Another receiver that is intriguing on this list is Jerry Jeudy. Jeudy played for the Denver Broncos last year. In the NFL, they are a rather dysfunctional team and known for mishandling the value of player personnel. Jeudy is a great player on a bad team. The quarterback play on the Denver Broncos last year was below average in 2023 and that resulted in less on field output for Jeudy. The result of the cosine similarity assessment means that Jeudy can be classified as a possible replacement for Smith and Brown. Jeudy would be less expensive than them because he was on a bad team last year and had less than average performance output.
DeVante Parker is a seasoned veteran, being drafted in 2015. Even though he was just added to the Philadelphia Eagles roster in free agency, someone with a similar output that the Eagles can replace him with that can possibly cost less than DeVante Parker’s salary is Michael Gallup. Having played on the Dallas Cowboys last year, he is now a free agent. If time goes by without Gallup being added to a team, the Eagles can offer him substantially less than what they did to Parker and have similar performance output.
Limitations:
There are some limitations in this dataset and analysis. This analysis can not take into account human factors associated with all sports. Players can be in slumps that affect their performance greatly. Severe injuries can be prolonging and affect them for years. Having been healthy, they would be a lot better than what they are on paper. In addition, the dataset only takes into account 2023 wide receiver stats. Explosive receivers that did not play well in 2023 are at a disadvantage in this analysis. Aforementioned, Cooper Kupp set many NFL records in previous years, but was dealing with injuries last year. He is an exceptional player, but it was not reflected well last year.
Github:
https://github.com/dhruvitpatel5/Final414Eagles/blob/main/final414.ipynb
