The Cavalier Daily
Serving the University Community Since 1890

Virginia Athletics leverages data via advanced analytic techniques

Football and women’s soccer lead the way in the use of sports analytics at the University

<p>Thomson (left) and Edwards (right) have both made major contributions to their respective sports, and they hope for similar data analytics roles to be created for other varsity teams at the University.</p>

Thomson (left) and Edwards (right) have both made major contributions to their respective sports, and they hope for similar data analytics roles to be created for other varsity teams at the University.

Since the novel and film “Moneyball” entered national discourse and advancements in technology enabled big data processing, sports franchises and fan bases have all taken an interest in data analytics and its potential role in improving player and team performance. The majority of research efforts and dollars spent come from professional sports teams who have more resources at their disposal. In fact, today, all major sports leagues have teams with in-house sports analytics teams dedicated to gaining an edge over the competition. With the majority of gains being made in professional sports, how have collegiate athletics responded?

Here at Virginia, the football and women’s soccer teams are pioneers in the collegiate landscape. Director of Football Analytics Matt Edwards and Eilidh Thomson, director of analytics and operations for women’s soccer, both saw an opportunity to leverage data on their respective teams and forged their own paths by creating analytics positions. Both directors witnessed the positive impacts that could come from data from their previous coaching positions and decided to develop a technical expertise for data analytics. With different experiences and backgrounds, the two directors are changing the way Virginia Athletics approaches decision-making and game preparation. 

Matt Edwards, Virginia football

Edwards serves as the current director of football analytics after serving his first three seasons at Virginia as a football analyst. He was formerly a tight end for BYU and a mathematics major. After graduating, he completed a masters in public administration and served as a graduate assistant for Coach Bronco Mendenhall at BYU. This connection eventually brought Edwards to Virginia football in 2017. 

Since joining the Virginia football staff, Edwards has enrolled himself in sports analytics classes and is currently pursuing a masters in data science at the University. Edwards describes Mendenhall as “a data-driven guy” who ultimately gave him a chance to introduce analytics to the Cavaliers’ football operations. 

Edwards’ current role has him generating weekly reports on two things — team performance and the opponent of the week. These reports are used for game preparation and in-game decision-making. The reports are built off models Edwards builds to analyze player and team performance and to predict certain outcomes from in-game scenarios. Player tracking data collected during practice and pro football focus data that houses game data on the Cavaliers and other opponents are the data sources used by Edwards.

Edwards uses every practice as a chance to collect traditional statistics on player performance such as completion percentage, total yards generated by each player and other similar metrics. However, the most useful information comes from the 40-plus wearables the team uses during practice. The data produced gives details about player location, exertion, explosiveness and more, which are all used to estimate expected points and evaluate player decision-making. Wearable data has also expanded analytics beyond improving performance. Edwards works with sports scientists on injury prevention using the data on player exertion and load. 

After collecting this data, Edwards primarily meets with different team staff members to run through each report. This information is then shared with the players at the coaches discretion. On game day, Edwards sits in the booth with a binder full of reports and hard output to provide coaches with data at a moment’s notice.

Virginia football is one of a small number of college football teams with a full-time analytics position, according to Edwards. Most other Division 1 football teams use data processed by a third party, which only includes standard, rudimentary game data. In this regard, the Cavaliers are ahead of the curve. 

Eilidh Thomson, Virginia women’s soccer

In an even more unlikely role, Thomson is the only director of analytics in NCAA women’s soccer. Similarly to Edwards, Thomson joined a very forward-thinking staff with Coach Steve Swanson at the helm. Swanson, associate coach Ron Raab and assistant coach Jaime Frias previously used video analysis to assess team and player performance and were immediately receptive to the addition of data analytic tools. After serving three years as volunteer assistant coach, Thomson created and filled the first-ever director of analytics position for Virginia women’s soccer.

“Data has become more influential in the game at higher levels,” Thomson said. 

However, unlike Edwards, Thomson has no real formal education or degree in sports analytics, mathematics or data science — rather, she is entirely self-taught in a multitude of different programming languages. In a sport like soccer with fewer sports analysts and researchers, Thomson uses her own experience and intuition to base her analysis on questions and insights she is interested in.

Thomson, similarly to Edwards, implements in-house data collection via GPS tracking data. During games and practices, players wear heart-rate monitors and a small GPS tracking device on the top of their spine that provides location and movement on the field. This data is then embedded into clips via SportsClips, a video analytics tool — Thomson finds that matching the data with film makes it more captivating and useful for both the coaches and players. The film can then be used to provide tactical and technical data on player performance and physical data to monitor player loads used for injury prevention. Finally, Thomson puts together match and season reports based on the film and physical data to provide the coaches and players with additional information.

Future of sports analytics at Virginia

In both football and women’s soccer, data analytics can help achieve that extra “one percent” — the difference between winning and losing a national or conference championship. Edwards’ and Thomson’s work has already led to overall better team performance, as evidenced by the fact that both teams have had favorable records since they joined.

“Analytics can critically shape the thought process and direction with which teams go,” Edwards said.

With Mendenhall at the helm and Edwards in the box, the Cavaliers have progressed from two wins in 2016 to nine wins in 2019. This past season Virginia fell to 6-6 overall but has remained very competitive in ACC play. Despite the underwhelming season, Edwards will use the offseason to take a deeper look at the previous season’s data and identify team weaknesses that can be addressed by the coaching staff ahead of next season. Data collected the past few seasons enables Edwards to make models of comparison and make better analyses and predictions to optimize team growth.

Under Swanson’s and Thomson’s guidance, the women’s soccer team has strung together multiple double-digit win seasons and NCAA Tournament appearances, including three Round-of-16 exits. On the surface, there is tangible evidence of improvement and success for both teams at the highest level of collegiate athletics.

In addition to football and women’s soccer, Virginia basketball also analyzes player performance via the Noah Shooting System, which uses sensors above the basket to perform shot analysis. Only a handful of other NCAA basketball programs use similar tools. Along with these three Virginia programs, a number of other varsity teams at Virginia have expressed interest in working with data. Edwards and Thomson have collaborated with other teams and assisted in expanding the use of data analytics to other teams. 

While data analysis can be highly effective, the opportunity cost for pursuing data analytics can be quite high, and the suspension of competition in the spring of 2020 due to the COVID-19 pandemic has exposed the fragility of many athletic departments’ budgets. However, given the positions carved out by Edwards and Thomson and the interest across Virginia Athletics, it is up to the athletic department to expand resources.

“It would be great to one day have a department within the athletic department dedicated to data,” Thomson said. 

Team-specific analytics is a step in the right direction, but both Edwards and Thomson have a vision of leveraging data analytics across the board via athletic department resources and support.

The University academic community has stepped up and initiated a number of initiatives to assist with and promote sports analytics. In the past few years, the University established the School of Data Science, where Edwards is currently pursuing his masters degree. The department of statistics has also created the Sports Analytics and Statistical Library, which is dedicated to sports analysis education and research. Additionally, the School of Engineering and Applied Science worked with the football team to build valuable models via yearly student capstone projects. These initiatives have resulted in student internships for a number of Virginia undergraduate and graduate students, and even led to two former interns landing jobs in the National Football League.

With so much support from the University along with the resumption of collegiate athletics, there is hope for more full-time analytics staff and resources in the near future. Based on my own participation with SASL and conversation with Edwards, collaboration is taking place between the academic community and the department of sports medicine, which deals with the healthcare of all Virginia student athletes. Given the positive trends in analytics across the University, Edwards and Thomson suggested the possible creation of a department of sports analytics dedicated to improving performance across Virginia Athletics. Furthermore, there could be a role for data in the recruiting process — however, it is hard to verify data from random, amateur sources and to find good data without collecting it yourself. 

In the future, look for sports analytics to continue to play a more significant role in Virginia Athletics as more teams use data to find that extra “one percent.”

Comments

Latest Podcast

The University’s Associate Vice Provost for Enrollment and Undergraduate Admission, Greg Roberts, provides listeners with an insight into how the University conducts admissions and the legal subtleties regarding the possible end to the consideration of legacy status.



https://open.spotify.com/episode/02ZWcF1RlqBj7CXLfA49xt