The tech behind The Batting Lab featured on NBC’s Today Show

It’s built on SAS® technology used to help kids understand how data is a part of their everyday lives. 

As highlighted in an NBC Today Show segment, SAS built a literal plate for kids to step up to with The Batting Lab. Created to improve kids’ understanding of data by improving their performance at bat, this AI experience combines statistics, machine learning and computer vision to analyze swings in real-time.  

Where does SAS technology come in? 

Thousands of swings were filmed of NC State University softball and baseball programs. Our team took that data and built a hidden Markov model using SAS® Econometrics, creating a 38-dimensional time series model of what an optimal swing looks like for an NCAA Division I player. 

In combination with the hardware in The Batting Lab, including a microphone, cameras, ball and weight sensors, the in-cage data is captured. SAS compares the youth player’s swing to those of elite college players to look for deviations, then delivers real-time feedback on how to improve. 

How are swings analyzed? 

1. Cameras and sensors capture all elements of the batter’s swing. With each frame, the person is identified alongside all their joints (shoulders, hips, knees, eyes, wrists, etc.). 
2. SAS processes the data and applies machine learning and business rules to create recommendations on swing improvements. It even makes it more like a game by scoring each element of the swing and comparing it to a data set of elite players. 
3. The cage then visualizes feedback as a recommendation for how the player can improve their swing. 
4. And the capture of data, analysis and data representations and recommendations back to the kids happens in seconds. 

How does the lab decide what recommendations to offer? 

Working with NC State University coaches, we identified 40 areas of feedback youth players commonly need improvement on and ranked them on a scale from 1 to 5. The 40 areas for improvement are evaluated for each swing along with the overall swing score and swing frame deviations. All this data is fed into a simple recommendation system to identify two pieces of feedback for the player. Feedback is prioritized by its criticality and how early in the swing it occurs. 

Once the feedback is determined, the system presents age-appropriate language to the player. Here are some examples: 

• Your feet should be shoulder-width apart. Spread them out a bit. 
• As your weight shifts back, move your hands so they are directly above your back foot and not higher than your back shoulder. 
• Try keeping your shoulders level while keeping your head and torso upright. 

This is the same technology some of our customers are using for applications like detecting flaws in products moving through a manufacturing production line, exponentially increasing the speed of processing lab tests, protecting endangered species like rhinos and koalas, getting ahead of treacherous winter road conditions, identifying deforestation in rain forests and monitoring cancer progression by analyzing CT scans.