The Potential of Sports Analytics to Revolutionize STEM Education

John Drazan
Apr 5, 2017 · 14 min read

Reflections on sports, science, and STEM education from one of the authors of the Grand Prize Winning Research Paper at SSAC17

As a basketball playing engineer, attending the MIT-Sloan Sports Analytics Conference (SSAC) has always been a dream of mine. Presenting a research paper, let alone winning the Best Paper award was a pipe-dream. When I was in college, in between basketball practice and physics classes, I would read Grantland and wonder how I could make my own heat maps a la Kirk Goldsberry. His work and the work of other luminaries in sports science and analytics (many of whom have presented at SSAC) provided me with a context and a motivation to persevere through difficult courses as an undergraduate in physics. As a basketball player, my coaches would often tell me something and I would ask “why?” While they didn’t always have satisfying answers, once I found sports science I was able to find them on my own. This changed my perception of science as a teenager. It went from a dry, academic topic in no way applicable to my own interests, to a useful tool that helped me better understand an activity that I enjoyed. Sports science presented me with an alternate perspective on what it means to be a scientist, one that was contrary to popular perception of a white dude in white lab coat mixing chemicals.

What makes someone a scientist? What do they study? Who can be a scientist? Pop culture provides the answer to these questions in the form of representations such as the hit TV show Big Bang Theory. Scientists are presented as quirky, (mostly male) geniuses who have trouble with social norms and navigating basic interpersonal relationships. They are BRILLIANT though, and they can achieve amazing, fantastic, feats for the betterment of society. Just don’t expect them to hold polite conversation or get the girl.

Science is easy, understanding people is hard…… :(

While I find these stereotypes to be annoying on a personal level as a practicing researcher, they are much more damaging to youth who are excluded from future STEM (Science, Technology, Engineering, and Math) careers because they don’t identify with these stereotypes. It has become common place for people to say “I just don’t get math” as if people who can do math have some sort of intrinsic ability that allows them to solve math problems. While there certainly are exceptions (The portrayal of Dr. Katherine Johnson in the movie Hidden Figures comes to mind), the vast majority of scientists, engineers, and other STEM professionals simply found something that interested them that required math or science and learned it along the way. For me, this interest was basketball and by extension sports science. Many people view math and science ability as an intrinsic aptitude, rather than a learned skill, because they are not provided with opportunities to make their own connections with STEM within their own life experiences.

With the benefit of time and study, I can look back and identify this disconnect as an endemic issue within traditional STEM education. Many students experience introductory STEM education as the study of existing scientific knowledge, not as the process of generating new knowledge. For students with familial role models within the STEM fields, these relationships can provide context and motivation to persevere through the “gate keepers” of the STEM disciplines. Additionally, as STEM is typically taught from the western perspective with mostly white, male luminaries, this may prevent females or students of African and Hispanic descent from identifying the preexisting STEM knowledge presented in class as part of their own heritage. This lack of representation may explain why female and underrepresented minority students often view STEM topics as being too difficult, not interesting, or inaccessible to them. This disconnect is not just depriving our society of future scientists, coders, and engineers with diverse perspectives and life experiences, it is also preventing social and economic mobility for the those in most need.

The recruitment of future STEM workers begins when youth are in their early teens. One of the biggest indicators of future study and employment within the STEM fields is participation in out of school STEM activities in middle and high school. Out of school STEM activities are crucial because students are allowed to generate STEM knowledge on their own rather than memorize the results of decades or centuries old science.

Typical out of school STEM activities range from robotics clubs and science Olympiad competitions to participation in research housed at local universities. While these programs are very beneficial for participants, they have difficulty recruiting students without a preexisting interest in STEM or familial support. Students disinterested in classic STEM tropes about lasers and robots are not going to sign up for these programs. It’s ironic that in the past decade, advances in technology, analytics, and computation have transformed almost every industry and activity into a STEM endeavor. Educational practice and popular perception of STEM have just not caught up with the times. Science and technology pervades everything we do, not just building robots. In effect, our society is trying to engage the youth of 2020 with STEM tropes from 1980.

The growing importance of STEM within sports is apparent by looking at the growth of SSAC itself. What started as a conference with 75 people in an MIT classroom has grown over the past decade to 4,000 attendees with panels headlined by sports executives, team owners, and leading journalists. Just perusing the finalists and posters in the research paper competition reveals the present and future applications of STEM and analytics in sports. Machine learning provides analysis and optimization of in-game strategy, player tracking allows for automated play classification for opponent scouting, and even the measurement of muscle fatigue allows for driver performance metrics in IndyCar. Although the old guard of sports like Charles Barkley and Goose Goosage are loathe to admit it, many aspects of pro sports are quickly becoming just as technologically advanced as any robotics lab. This leaves us with a two questions regarding sports analytics and youth sports.

  1. How can we leverage the burgeoning field of sports analytics to train youth players?
  2. How can we use youth sports to provide an authentic introduction to youth who are excluded from the traditional STEM pipeline?

Answering these questions will enable our society to produce better athletes while also broadening access to STEM education among our most under-served youth.

At first blush, it would seem that the growing “techiness”of sport has been well communicated and accessible to everyone. The revolution, quite literally, has been televised. At the same time I was reading every heat map article Kirk Goldsberry wrote on Grantland, I was also watching ESPN Sport Science clips on YouTube. The question that was always in the back of my head while I vociferously consumed sport science content was “How can I measure this myself so I can get better?” The answer when I was in high school was “well I can’t.” So I went and studied physics in college while playing basketball. In college I made a primitive gait biomechanics lab as a senior project so the answer became “I kinda know how, but I’m not there yet.” So I went to graduate school to get a doctorate in biomedical engineering. The answer now is “Yes I can! But its a few years too late……”

Data collection from my undergraduate biomechanics lab.

It took me nine years(!) of post-high school education to reach a level of expertise with cutting edge biomechanics technology. At 27, after being out of the NCAA for five years, my competitive basketball playing career is at an end outside of men’s leagues and pick up. Luckily for me, I am now safely ensconced in my career path in STEM so I can use engineering to feed myself, however, it would have been been nice to use this when I was playing. I realize now that authentic applications of sports science are not accessible to the majority of the population, especially those from marginalized communities.

While I didn’t know it at the time, there was a McDonald’s All American basketball player in attendance at SSAC17 by the name of Mohamed Bamba. In a recent interview, he had this to say about his experiences at Sloan:

I learned about the game of basketball from an analytics view. We did an entire seminar with a panel on how to guard a pick-and-roll. I remember getting up and asking a question, ‘Does this translate at all to the high school level?’ They didn’t really have an answer for it.

This captures both the allure and limitations of analytics within youth sports. A top 5 basketball recruit in the nation paid hundreds of dollars to attend SSAC to learn about analytics, yet he was left with no way to apply what he learned within his own game. The technological barriers to entry for sports-science not only makes it difficult to use as a STEM outreach tool, but it also prevents youth coaches and players from using it as a tool for improvement.

For the all the limitations of the traditional STEM pipeline, it is very good at scaffolding scientific material within students existing knowledge. Lego Mindstorms provides an age-appropriate introduction to robotics, the issue is that a lot of kids aren’t that motivated to build robots during their free time. Many more kids love playing or watching sports. If we can similarly scaffold sports science and analytics as a tool for improvement within youth sports, we can authentically introduce STEM to a broad set of youth who are interested in sports at the outset rather than STEM. In this manner students without a preexisting interest in STEM can be engaged in STEM enrichment. Just as students in robotics clubs are motivated to learn STEM to build the best robot and win competitions, basketball players could be motivated to learn STEM to use sports science to become better and get a scholarship. Youth players just need an accessible, authentic set of scientific and analytical tools that enables them to study their own games.

If we can make basketball an accessible venue for youth to learn how to use STEM and analytics as a training tool, we can turn STEM knowledge into a competitive advantage that players can use for improvement. This is the motivation for our project titled “From Sports to Science: Using Basketball Analytics to Broaden the Appeal of Math and Science Among Youth” This motivation gives our project a very different focus than previous papers presented at SSAC. Most of the previous finalists and winners developed analytical tools that provided new insights into sports performance. Our project is different. Rather than developing a complex analytical model that explains sports, we instead created a tool for youth sports programs using simplified shooting analytics within basketball.

Sports analytics has huge potential for youth athletes, but its difficult to apply without the proper tools or background. Just ask our friend Mohamed Bamba, or if you have a time machine, me, 10 years ago. Fortunately, the recent advances in basketball analytics are built upon a legacy of statistical analysis of basketball that was developed before the advent of advanced tracking technologies and analytics. Valuable performance metrics such as shooting percentage and scoring efficiency can be determined by collecting data with a pen and paper and calculated using simple arithmetic. These simple metrics allow players to utilize basic mathematical concepts such as fractions, decimals, and percentages combined with pencil and paper data collection authentically within basketball to understand their own performance. While collecting data with a video camera and player tracking may be slightly more efficient and lead to more detailed analysis, this prevents players from understanding the process and ultimately internalizing the information. It also makes it significantly more expensive. Our methods just require a pad of paper, a pencil, and kids ready to work.

For our paper we created a program that provides youth players and coaches with an important training tool by combining these accessible analytical basketball metrics with modern graphical representations of data such as heat maps. These heat maps effectively communicate spatial shooting data that shows players where they need to improve as shooters, where they should be shooting from, and how they can be effective players. The athletes collected their own data with a pen and clip board so that they understood the entire process, from data collection all the way to data analysis. After data was analyzed by our heat map program, we gave athletes their own heat maps so that we could provide personalized coaching and analysis of the results.

We used this shooting analytics program as part of series of free shooting clinics that we deployed around our area. We ended up working with almost 100 student athletes across six different clinics. Participants were recruited to participate in a basketball clinic, not a STEM activity. We gave the students a survey before and after participation to see how their perceptions changed about analytics, STEM, and basketball as a result of participating in this activity. Participants reported an increased understanding of their own game, an increased interest and understanding of analytics within basketball, and we even showed an increased interest and enjoyment of STEM outside of school.

Student athletes viewing results at a clinic with co-author Amy Loya

In effect, we were able to extend the many of the benefits of modern basketball analytics to the general population who otherwise don’t have the technical resources or expertise to participate in the modern analytics revolution. We were able to show that using easy-to-understand sports analytics tools not only increased athlete knowledge about their own performance, but it also increased their appreciation for analytics and their interest in STEM fields. Especially noteworthy was the fact that we attracted students interested in basketball rather than STEM, thus broadening the reach of STEM enrichment programs. We had created a validated tool for basketball analytics that my teenage self would have loved. My team and our student-athletes certainly loved it. The only question left was whether or not the research paper committee for SSAC would love it. In the back of my mind, I wondered how our simplified approach for sports analytics for youth training and education would be received at the Mecca of Analytics.

With these results in hand we submitted our abstract to the SSAC committee on the “Business of Sports” track. Our abstract led to an invitation to submit a full manuscript and in the end we were selected as one of the top eight submissions. After my 20 minute presentation in the semifinals I got an incredible amount of support and positive feedback from people in the NBA, sports industry, other academics, and a ton of students. Everyone was extremely excited about not only the social and educational impact of our work, but also its potential to improve youth sports. When I found out about making it to the research final presentations, I returned to my room and began practicing my final 10 minute presentation. I dove back into our paper to make sure that I could respond to any technical question regarding our work and methods.

I was the first presentation in the finals and I stood facing a growing crowd. The judges for the finals were Nate Silver of 538, Shelley Pisarra of Wasserman, and Kirk Goldsberry formerly of Grantland and Harvard, now of the San Antonio Spurs. Nate was a little late so I snapped a picture of the crowd from the podium.

I think that I ended up going about 45 seconds over my allotted time. I fielded a question from Shelley about transitioning students from the basketball side of the program to the STEM fields. I was able to give examples of six of my athletes who are now studying STEM or health care topics in college. Kirk asked me a question about the age ranges that I could work with, and I responded with the demographics from the paper, grades 7th to 12th grade while describing how I changed the program for each age group. Finally I turned to Nate and I mentally reviewed my statistical model assumptions while he arranged his thoughts. His line of questioning was entirely fair, however completely unexpected.

This seems like really important research, and I would like to underscore the point that lack of representation is a problem for STEM fields[…]The question I would ask to the group is this: is this a sports paper or is this an education paper? I want to give you a chance to address the criticism that this is a sports research paper competition and this is not a paper about sports.

To say that I was surprised by the question would be an understatement. I had expected a tough question, however calling into question the relevancy of my project to the overall competition was unexpected. In the moment, I responded by talking about the importance of communicating technical information between analytics staff, coaches, and players and how our work shows how to do so most effectively. Nate nodded and I returned to my seat. I ran the question through my head for the rest of the day and had seemingly a million better responses by the time the Alpha awards came around.

As the subtitle to this post already told you, we ended up winning the competition so my response to Nate Silver’s question must have been sufficient. While I wasn’t too happy about the question at the time, looking back I am grateful to Mr. Silver because his question addressed the heart of our work. It was one of the main reasons why I decided to author this post. At the end of the day, if our approach is going to become popular within youth sports programs across the country, it has to be a useful tool for basketball players. Our data shows that it does. By enabling youth players to explore and understand their own performance using STEM, we are increasing that athletes ability to evaluate their shooting ability. The program is freely available online so anyone can use it. The potential reach of our methods are huge.

To talk about just the basketball training applications of this work is to sell it short however. There is a serious issue with how we provide youth with opportunities in the STEM careers. As the value of STEM degrees has increased due to economic forces, the under representation of certain groups in the STEM fields is now a matter of equality. The exclusion of females, racial minorities, and impoverished youth from STEM careers limits the social and economic mobility of the most marginalized while simultaneously depriving the STEM fields of diversity of thought and perspective.

Over a beer after the conference, one NBA analytics staff member told me “Analytics is a zero sum game. If someone uses analytics to get better, they are winning at the expense of another team. ” While this is certainly true within a sports league, our work shows that this analysis is flawed. The impact of sports analytics extends beyond wins and losses. It even extends beyond championships. Sports has a special place in the American psyche. Sports fandom crosses many racial, economic, and social lines, giving it unmatched reach into our society. The potential of sports analytics to revolutionize popular perception of science is real, and its potential for positive impact on STEM education for all youth is too great to ignore.

John Drazan

Twitter: @Sports2STEM



Albany, NY

Update: MIT-Sloan just posted videos from the competition. Here is the first presentation.

Education Reform

Writings on changing the status quo.