Analyzing NFL Player Injuries with Python

By Shuwen Li

With Greg Page

The National Football League (NFL) has recently begun to enforce rule changes in order to reduce player injuries. In particular, the NFL rule changes have been designed to reduce concussions, as well as career-threatening injuries to the knee area, such as Anterior Cruciate Ligament (ACL) / Medial Collateral Ligament (MCL) tears.

The NFL modified kickoff rules after reviewing data indicating that players’ concussion risk on kickoffs was nearly four times greater than their risk of a concussion on running or passing plays. The league also clarified penalties for improper equipment usage, such as players wearing unapproved helmets.

In the following report, I have analyzed three datasets in order to more comprehensively understand NFL player injuries.

● Injury Record: contains information on 105 injuries that occurred during regular season games over the two seasons, including details regarding the number of days that players missed due to injury. This dataset does not include head injuries, so concussion data was not included here.

● Play List: This dataset contains the details for 267,005 plays. This dataset includes information about player injuries, as well as general game details such as weather and field type.

● Player Track Data: This dataset includes details about particular players on the field, noting their location, orientation, speed, and direction during specific plays.

To start off, grouped injuries by body part, using the groupby() function in pandas, and then created a barplot to compare their count values. As the plot below indicates, the most common injuries occur to the knees and ankles.

Next, took a closer look at knee injuries in particular, to compare the incidence of ACL tears and MCL tears. Generally speaking, an ACL tear is more serious than an MCL tear, and may require surgery to repair. As indicated by the results shown below, MCL tears were more common during this three-year period.


Next, separated the injuries by field type. In the graph below, natural grass surfaces are represented on the left, with synthetic (artificial turf) surfaces on the right. This graph demonstrates that knee injuries occurred in similar numbers across both field types. Ankle and toe injuries occurred more commonly on synthetic fields, whereas foot injuries were more common on natural grass.

JC Tretter, the NFL Players’ Assocation (NFLPA) President, has strongly advocated for the use of natural playing surfaces, citing the additional strain on players’ bodies caused by artificial turf. Based on NFL injury data collected from 2012 to 2018, not only was the contact injury rate higher during practices and games held on artificial turf, but players also consistently experienced a much higher rate of non-contact lower extremity injuries on turf compared to natural surfaces.

Using the dataset with per-player position tracking for 250 players, across two seasons, I compared the lower body injury rates on natural and synthetic surfaces, with results shown below:

Even data listed could be bias, dataset hardly be absolutely accurate, those historical data recorded indicates that players are possibly to get injury on synthetic type. Furthermore, to support this opinion, I will implement Chi-square hypothesis.

Null Hypothesis H0: play injury, no different between artificial and nature turf

contingency table

A p-value of 0.73 is higher than the widely accepted maximum value of 0.05 which H0 can be accepted. The data does not strictly support the idea that players with more plays on synthetic turf have a higher injury rate.

Even so, this result doesn’t mean that the effect of playing on synthetic turf should be ignored, we only generate those 250 players as test simple. Actually, during practices and games from NFL 2012–2018 gathered data reports that natural grass fields provide a lower risk for injuries compared to artificial surfaces. Also, from physics side, football players put force and rub onto the playfields, grass will eventually release the cleat prior to reaching an injurious load. It should be considered that NFLPA asking teams to change all fields to natural grass.

Of course, other factors may also play important roles in predicting NFL player injuries. Among these are:

● Stadium type

● Weather

● Player Position

● Injury history

● Player conditioning

Still, the available data certainly supports the claim from the NFLPA — synthetic field surfaces are more likely to be associated with player injuries. The NFL should take this information into account when designing new stadiums, or when refurbishing existing ones.


# Knee Injuries, Regular Season Games; Source from: IQVIA