What is the NFL going to do with the 'big data' collected?
In order to find out a few of the possibilities, one must continue to read below. The possibilities are endless, although, the initial reasons are restrictive.
What Is Big Data In Football?
When I read the title of the article on the website "Scientific Computing" titled "The NFL Joins The Data Revolution in Sports", the first question that came to mind is:
What data are they collecting that is not already being collected?
I was confused by the title of the article, since, one would think that a huge organization like the NFL would already have an enormous amount of data. Think about the gambling industry across the world and their profits on sports. One would imagine that big data has played a significant role already in generating an enormous amount of profits from big data. Evidently not. Hard to believe.
According to the article mentioned, the NFL is just entering the field of "Big Data":
In some potentially game-changing news for the way we understand professional football, the National Football League began the 2016 preseason by placing tracking sensors in its footballs for the first time. The chips are also in balls used in Thursday night games.Over the past decade, we’ve seen an explosion in data analytics in sports, particularly on the professional level. Technological advances in cameras and sensors have allowed teams, media and fans to gain insight into a bunch of previously gray areas of sport performance, such as the National Basketball Association’s use of SportVU to track every bit of player and ball movement on the floor.The concept of integrating numbers and analysis into scouting, training and coaching isn’t new. But access to powerful hardware and software has greatly increased the quality and quantity of available data. A nearly insatiable appetite for data on sports has created a sports analytics market that is set to grow from the millions to the multiple billions of dollars over the next few years.
The amount of data generated during each game would be enormous. By keeping the sensor limited to the football and possibly the sidelines, the data generated would be reduced too. Although, with a reduction of data flowing in from the game, the accuracy of the plays suffer too. The author mentions that the next step would be to incorporate sensors into the players 'shoulder pads' - which would increase the data stream coming in.
Overall, the practice would be transformative to the entire industry. I wonder how that would change the challenges that referees face during the game. Currently, during a challenge, the play is reviewed on a closed circuit screen available to the referee and officials only. With the rise of sensors, now the game can be analyzed by each team in real time. Although, the technology is not distributed in real time yet.
Any avenue of improvement that the coaching staff can incorporate into the teams training regimen would be greatly sought after. Currently, teams are exploring both game simulators and drone coverage of their practices to improve overall flow. The incorporation of data from the NFL offers two great aspects of improvement:
Ideally, data from ball trackers or shoulder pad trackers could serve two purposes for the NFL. First, it can help teams understand player movement and the flow of play more completely, providing coaches a greater understanding on how players are physically performing during plays, and allowing for input from coaches to players on how to fix their technique to increase efficiency or limit exposure to injury, possibly leading to more efficient training and practice.Second, the data can be used by the league’s media partners, and perhaps its fans, to further explain the game to audiences, particularly on television. By tracking player movement digitally, clearer representations of what makes individual football plays succeed (or fail) can be provided. These data also allow media to break down individual physical accomplishments, such as extraordinary bursts of speed by wide receivers.The NFL’s plan to release tracking data within 24 hours of a game’s end points to a future in the league where hard data on player and ball movement are integrated into the daily strategic calculations of each coaching staff. This will likely create a rush to innovation within NFL coaching, as each staff grapples with what will likely be a huge amount of data every week, trying to come up with best practices and analytical methods for evaluating and using that data constructively.
Of course, generating a tremendous amount of data means that the NFL along with individual teams that participate need to have the technological infrastructure (computing power, data scientists, etc.) to make meaningful use of the data coming into the organization. This requires both technology and scientists to handle that technology in a fruitful manner.
That means scientists will be taken away from professional fields in which they were trained to contribute. Is this good?
NFL Data Science Improves Science Indirectly
There are a tremendous amount of scientists who are interested in sports. At least, that is my impression after going through the university system in a science driven field -- through an advanced degree program. The prospect of losing a scientist to the NFL organization at first sight might seem unethical. Scientists should stick within their field (discipline) right?
Not necessarily. There might be many benefits by losing data scientist to the NFL. First, the scientist working for the NFL will inevitably have a appropriate infrastructure to handle the large amounts of data coming in. In science, funding is scarce and often sought out among many research groups.
I have always maintained that in order to improve the funding for science, we need the entertainment industry and the sports industry to get involved (financially and technologically) to boost the ability of science. Why? Not all great ideas come from working on science problems in science.
What do I mean by this last statement?
A famous story about the world famous physicist Albert Einstein revolves around generating his best ideas while shaving. Successful people will often tell stories of ideas which have been generated about their business while performing outside work or tasks. The shower or shaving are just two.
Additionally, while performing a job outside a given field, a scientist may gain insight into the problems within their field. This methodology is sometimes referred to as "thinking outside the box." By tackling problems associated with dealing with large data sets like players in a game, other problems might be tackled using different algorithms. Can you think of any? I can.
One such problem is tracking people in real time in a city and finding potential threats (WM -- chemical and biological weapons, etc.). Sifting through the data to find meaningful answers might improve the governments ability to sift through data to find a threat. Although, the funding opportunities to develop an algorithm or simulation might be too costly on part of the city. Therefore, having organizations such as the sports organizations tackling the data regarding player movement within a given region (on field inside a stadium) will inevitably improve our ability to detect a threat.
As most of us know, the entertainment industry is rich in funding and not at a loss for funding such interesting projects. Alternatively, new algorithms will be made (which are proprietary for the NFL) to tackle the issue of analyzing real-time data. But the inherent thinking or structure of mining the data is what is critical. After that is known, then an algorithm could be changed to achieve that specific problem. This prospect offers a great future to science and society in the future.
The correlations which will arise as a result of data mining real-time player information have yet to be realized. By the descriptions in the cited article above, we are just at the tip of the iceberg in terms of finding relationships within such data sets. Additionally, no one knows the benefit or adverse effect the data mining will have on both the gaming (gambling industry) and the NFL organization.
Hopefully, out of such data mining algorithms, safer players (with less injuries, etc.) will result. Science will inevitably benefit out of the data mining processes that are developed. I have no doubt about that. Scientists are interested in sports and already use the industry to approach problems in science. Even if progress is made on the initial thought process of how to find correlations in the data, I believe that meaningful results will arise from the exercise. Initial findings suggest that this is the case. Although, as I mentioned, we are just at the 'tip of the iceberg' in the process. Stay tuned!
Until next time, Have a great day!