Skip to Main Content

Introduction to AI Through Sport

Since 2023, Stats Perform has partnered with Student Freedom Initiative and Morehouse College to deliver an Introduction to AI Through Sport course, which is designed to share insight into the application of artificial intelligence tools and techniques to analyse and predict performance across various sports. 

The course uses 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.  

Additionally, students also learn how to visualise granular data, as well as create interactive animations and dashboards to highlight various behaviours in sports. No prior knowledge of AI and machine learning is required to enrol.  

The course’s various lecturers, which include Stats Perform Chief Scientist Patrick Lucey and other members of Stats Perform’s AI team, use examples from our basketball and football (soccer) datasets and interactive tutorials with working code, with students also benefitting from access to these resources if they want to get their hands dirty.  

The goal of the course is to give students a basic understanding on the value of data and how AI maximizes the use of the data, and how it powers everything in the data ecosystem by utilising machine learning (ML), computer vision (CV) and large language modelling (LLMs) – using sport at the vehicle to learn.  

At the completion of this course, students should have an understanding of the basic concepts of AI and what it can and cannot do – i.e., the “why”. An additional goal of this course is to make students both data and AI literate, whilst emphasising the need to know the basic skills of utilising spreadsheets (Excel) as well as Python. 

The course, which runs from August through to mid-December, is broken into four parts with students required to deliver homework assignments and a final project, which all contribute to the final course grade. Students also sit a final exam prior to Christmas, which will account for 20% of their final grade.

LEARN MORE ABOUT PAST COURSES

Course Outline 

Part 1: Why Data Analysis and Visualisation is Required in the Age of AI

In part one, we give an overview of AI in Sport: the history of sports analytics and the reason why AI with sports data is necessary.  

  • We connect sports analytics to what occurs in business analytics, as well as provide an overview of the types of sports data available and how it is collected.
  • We provide a framework of the types of applications sports data drive and give an overview of what AI, Computer Vision (CV) and Machine Learning (ML) is and how they interact.
  • We highlight the current performance of AI Assistants, and how having the basic skills of exploring and visualising data is required due to the hallucinations that occur from AI Assistants.   
  • Using box score basketball data, we will do initial data-discovery and create the base metrics which are used throughout basketball (4 Factors and Efficiency Metrics) to teach the basics of Excel. 
  • At the conclusion of part one, we will demonstrate how we can create player metrics as well as visualise shot data in basketball, both using Python, and the various tools and methods to successfully give meaningful visualisations.

Part 2: Introduction to Machine Learning  

Using both basketball and football data, we will explore how to create predictive metrics as well as descriptive metrics. Specifically, in this part of the course, we will cover:   

  • Using win-probability as an example, show why machine learning models (using supervised learning techniques) are required to answer important questions in sport, such as “who is going to win” within a game. 
  • We will also introduce the concept of “counter-factual analysis” or “what-if” questions which can measure the impact of decisions during a game, such as when to take a time-out. 
  • Additionally, we will cover “Monte-Carlo” simulation, which can enable forecasts for where teams will finish at the end of the season using team power rankings where we utilise the ELO metric. 
  • Finally, we will also show you how to create your own expected goal value model in football, and how to evaluate it. 

Part 3: Introduction to Computer Vision and Player Tracking Data  

Computer Vision and tracking data have been a staple of sporting analysis over the last decade.  

Even though this is an advanced topic, we will just give a high-level overview of how it works, how it is used and the future of this technology in sport.  


Part 4: ChatGPT and LLMs  

In this part of the course, we will highlight the recent advances in AI, specifically around ChatGPT and Large Language Models. Students should have an intuition on what they are and how they work.  

In addition, we will highlight how they can be utilised in sport and will show students how to build their own.   

A crowded basketball stadium

Are You Ready To Build Magic With Us?

Get in touch today to find out how we can work together to grow audiences, talk to them more, increase and open up revenue streams and achieve on-field success.

get started today