Ahead of the Fall 2023 semester, we announced a long-term strategic partnership with Student Freedom Initiative (SFI) and Atlanta University Center Consortium (AUCC) to create and deliver a for-credit AI in Basketball course, taught by Chief Scientist, Dr. Patrick Lucey.
Given that basketball has been at the forefront of advancements in AI over the last decade or so, and Stats Perform pioneering much of the technology innovation with products like SportVU and AutoStats, as well as several prediction models and automating insights via natural language generation, this partnership was a logical one.
The course aimed to give STEM-majoring students at AUCC member institutions like Morehouse College, Spelman College, and Clark Atlanta University a basic understanding on the value of data in basketball and how artificial intelligence maximizes its use, and how data power everything in the sporting landscape.
“This course offers more than just a look into the applications of AI in sports. It also provides technology education, as well as internship and career opportunities for participating students,” said Robert F. Smith, Chairman of Student Freedom Initiative, and Founder, Chairman and CEO of Vista Equity Partners. “But most importantly, it offers the ability for those from communities underrepresented in STEM to learn more about a cutting-edge and evolving field.”
Throughout the course, taught on Morehouse’s campus, enrolled students were introduced to machine learning, computer vision, and large language models, and got a hands-on look at data analysis, data visualization, and player ratings.
Learning opportunities extended beyond the classroom when Dr. Lucey and the students attended HBCU Night hosted by the WNBA’s Atlanta Dream, seeing first-hand how what they’ve been taught throughout the semester can be applied in real-time.
Carl Mergele, CEO of Stats Perform, also served as a guest lecturer for a class, imparting his knowledge on the industry, and the importance of the course on increasing diversity in the workforce and building a talent pipeline.
Looking Back on the First Edition of the Pioneering AI in Basketball Course
Meet Ramiah Curry and Chavier McDaniel, students of the AI in Basketball course. Entering their senior year at Morehouse, Ramiah is majoring in Computer Science and looks to pursue a career in AI, while Chavier double majors in Mathematics and Physics, aspiring to take on the aviation industry.
Following the conclusion of the course, we sat down with the two to get a better understanding of what they learned and the outcome of their final project, creating a Chatbot using AutoStats data and a large language model.
Below are a few highlights from the interview.
From the AI in Basketball course you did over the fall, what do you understand about AI that you didn’t before?
Prior to this course, we had a visionary idea centered around leveraging the Large Language Model (LLM), ChatGPT to be a conversational agent, to help with code generation, or craft engaging stories.
Before diving into this class, we lacked a comprehensive understanding of the inner workings of AI in processing and utilizing data for predictions. Now, we’ve gained insights into properly sorting data, ensuring clarity and conciseness through charts and responsive feedback mechanisms, using large language models like ChatGPT. Understanding that not all data is equally valuable, we’ve learned effective strategies to sift through information, enhancing the precision of AI-driven decisions.
What aspects of the course did you enjoy?
Ramiah: I particularly enjoyed diving into the realm of computer vision and visualization. Along with topics of annotation and computer vision, this class gave me a way to enrich my knowledge of AI in other realms such as using graphs to better visualize the data. The interactive Google Colabs we used provided an excellent platform to apply these concepts. Utilizing these resources in class with a data scientist, Dr. Lucey gave an exciting trek into this world with helpful guides to build my own tools. I was able to build useful, practical knowledge while collaborating with peers.
Chavier: The parts of the course that really got me excited were the LLM and the computer vision sections. The LLMs (that’s Large Language Models, if we’re keeping it casual) were a blast. Last summer, as a personal project (such as AI storytellers/ book writers), I dove deep into Chat GPT, wanting to unravel its full potential and figure out how to use LLMs like a pro. The course was perfect timing to understand the inner workings behind these language models even more. And then there’s the computer vision part. It aligns perfectly with my goal of creating an aircraft infused with AI. Computer vision is like the secret sauce in making my dream aircraft a reality. It’s fascinating to see how these pieces fit together, you know?
In the final project for the course, you were asked to create a Chatbot using AutoStats data and an LLM model. Can you describe the Chatbot that you created?
Our Chatbot channeled the unmistakable flair of none other than the legendary Charles Barkley. We modified this cutting-edge language model to sift through AutoStats data from Stats Perform, offering a conversational and friendly dialogue to assist in player analysis and comparisons. This Chatbot serves as your personal Basketball GM assistant, delving into the depths of AutoStats draft data, box scores, and consensus to provide you with the lowdown on your chosen player. So, get ready for a slam dunk of information delivered in a style that captures the essence of the one and only Charles Barkley, making your sports blog experience a truly engaging and insightful journey into the world of basketball analysis.
As part of your final project, you were also asked to do specific analysis of a player. Which player did you analyze, and what interesting things did you find about him?
So, let’s talk about the man, the myth, Kostas Antetokounmpo – the Greek-Nigerian sensation and brother to the mighty Giannis. As we delved into the AutoStats data, we uncovered some intriguing nuggets about this forward. Known for his shot-blocking prowess and smooth transitions to conversions, Kostas caught our eye. When it comes to offensive prowess, he’s dancing in step with Mo Bamba, Jarred Vanderbilt, and Ray Spalding. Now, our Chatbot tossed him the 60th spot in the ranking given our data, mirroring his NBA career standing. This standing was based on many statistical factors, such as his strength in offensive adjusted rebounds and nearly 100th percentile ranking in turnover percentage, along with his weakness in contested 3-point percent defended. But during my deep dive into biases seen in the data, I stumbled upon a tough break – a tibia fracture in Greece sidelined him for the entirety of his first season at the University of Dayton.
You can read the complete interview, including a more in-depth explanation on their final project, its findings, and what the future holds for Ramiah and Chavier, here.
Ramiah and his classmates Noble Kemp, John Jackson, Isaiah Wimbush, Ronny Kiprono, and Amir Harris took part in our eight-week summer internship program, where the six Morehouse students got first-hand experience working alongside our Artificial Intelligence team.
During the program, the student interns were joined by Dr. Lucey, Carl Mergele, Elizabeth Cutri, General Counsel and Chief People Officer, and Mike Perez, Chief Operating Officer, for a dinner in Atlanta.
A New School Year, A New Chapter
Following the success of Fall 2023, Dr. Lucey has returned to campus for Fall 2024 to deliver an AI in Sport course, where his teachings will cover the integration and application of deep data and artificial intelligence across the sports ecosystem.
Similar to most businesses, the last few decades have seen sports organizations begin utilizing data and analytics to objectively measure performance and enhance decision-making, a central theme of Michael Lewis’s book, Moneyball. Now, organizations are looking to more granular data sources like spatial locations and ball-tracking data to get an edge over their competitors.
At the completion of the course, students should have a general understanding of key concepts in AI and what it can and cannot do when applied to the sporting landscape. Through each lecture, Dr. Lucey aims to inspire students to continue studying AI and prepare them for jobs in the emerging field.
“The value proposition of this course is not to create more data scientists in sport, although it is a very good starting point, but to use sport as the vehicle to learn how data and AI can be used to ‘measure the immeasurable’ via new insights and technology that could not be measured before,” said Dr. Lucey.
On April 3rd, Dr. Lucey was joined by Jonathan Tellez, Director of DEI and Employee Development, and Ysabel Gonzalez-Rico, Data Analyst, on Morehouse’s campus to promote the upcoming AI in Sport course. The team met with the Junior and Senior Computer Science Seminar to share more about the opportunity and give a glimpse at the how and why of data’s utilization and implementation across sport.
In addition to meeting with prospective students, Ysabel and Jonathan hosted a career booth at The Association of Computer Science Departments at Minority Institutions (ADMI), meeting with students and faculty from minor-serving institutions to discuss opportunities available at Stats Perform. They also had the opportunity to meet with students from the previous AI in Basketball course and see them present their final project Chatbot to conference attendees.
