Carnegie Mellon University

Carnegie Mellon Sports Analytics Colloquia

sports analytics graphic

Data-informed sports analysis with experts from ESPN, the NFL, the NHL, and more.

The Carnegie Mellon Sports Analytics Colloquia program explores how advances in technology, the availability of accurate, granular data, and wider acceptance of sports analytics has transformed how we approach analyzing players, teams, and games. Teams, media organizations, and fans have embraced data-informed methods of analysis as a way to confirm beliefs, debunk myths, alter strategies, and raise new questions. The legalization of sports betting has further increased the popularity of sports analytics and has even led to improvements to data integrity.

In this series of workshops, we discuss fundamental questions in sports, and how data and analytics can be used to help teams, media organizations, and fans answer them. Coding experience is not required. Resources, including example code and data sets, will be provided.

Program Details

Dates & Times

  • Dates TBA


  • online 


  • $550 per session or $2,200 for all five sessions

Carnegie Mellon alumni can receive a discount on program costs. When registering, select “Yes” from the drop-down menu under the “Are you an alumnus of Carnegie Mellon?” question on the form.

There also are a limited number of 50 percent discount student spots and a limited number of needs-based scholarships available. When registering, select “Yes” from the drop-down menu under the “Are you currently a student?” question on the form. You will then be contacted with application information. 

Program Session Details

Communication, Visualization, and the Data Science Workflow in Sports Analytics

Guest speaker: Christie Aschwanden, author of "Good to Go: What the Athlete in All of Us Can Learn from the Strange Science of Recovery"

Communication is an essential part of several stages throughout the data science workflow. Visualization is a powerful and effective means of communicating key results and observations, as well as a vital tool for data management and data exploration. In this session, we will discuss:

  • Some of the most important data science problems faced in sports analytics; the main goals when solving these problems; and how those goals differ among teams, media organizations, and other sports organizations.
  • Best practices for creating data visualizations that are easy to understand at a quick glance and highlight the key information you want to communicate while providing appropriate context.
  • How these best practices change when you are communicating to different audiences, such as team management, the general public, or fellow analysts.

Randomness and Uncertainty in Sports

Guest speaker: Brian J. Burke, Senior Sports Analytics Specialist, ESPN

Sports with relatively few scoring events — such as football, hockey, and soccer — suffer from small sample size issues when using goals or points to evaluate teams or players. Even for some sports with higher scoring rates — such as basketball and some statistics (e.g. opponent 3-pt FG percentages) — can be subject to randomness and take a long time to stabilize. These properties can be problematic when attempting to predict future performance of teams and players. Expected points and expected goals models (and the player and team metrics derived from them) help to alleviate some of these issues while serving as the foundation for many types of analysis. In this session, we will discuss:

  • Randomness and uncertainty in game events, team statistics, and player statistics.
  • Probability distributions and their role in sports analytics.
  • Expected points (basketball) and expected goals (hockey, soccer, and lacrosse) models which estimate the probability of shot success in those sports.
  • Expected points models in NFL and college football, and example applications.
  • Team and player statistics that can be derived from those models, and how their stability and predictive performance compares to traditional statistics.

Team Ratings and Predicting Game Outcomes

Guest speakers: Paul Sabin, Senior Sports Analytics Specialist, ESPN; and Lauren Poe, Associate Sports Analytics Developer, ESPN

An important step in evaluating or predicting team performance, or the outcomes of games and seasons in team sports, is developing an overall rating for every team in a league. In most leagues, it is important to account for the differences in opponents, schedule difficulty, and other factors that may affect a team’s game results or season standings. In this session, we will discuss:

  • Factors to consider when developing team ratings, including what information to include, how the outputs will be used, and how the outputs will be expressed.
  • Regression-based approaches to modeling team offensive and defensive performance that account for opponents and other factors.
  • Using team ratings to predict game outcomes and season win totals, and using probability distributions to estimate the range of possible values and the chance that each occur.

Player Ratings and Projecting Player Performance

Guest speaker: Jack Birch, Ph.D., Professional Scouting Consultant, Winnipeg Jets, NHL

Single-number metrics that estimate overall player performance are useful for giving a quick view of the player’s contributions to their team. With these metrics, as with team ratings, it is important to account for opponents and other factors that may affect a player’s individual statistics or the performance of the player’s team. In team sports, it is even more important to control for the quality of a player’s teammates, and in many cases, techniques similar to those used for developing team ratings can be used with players as well. In this session, we will discuss:

  • Adjusted plus-minus models for players in basketball, hockey, football, soccer, and lacrosse, which use regression techniques to account for teammates, opponents, and other factors.
  • The idea of replacement level players and points, goals, or wins above replacement metrics.
  • Using player metrics along with age data to create age curves and project future performance of players.
  • How player metrics can improve team ratings and game predictions, especially in the COVID-19 era.

Analyzing Gameplay with Player-Tracking Data

Guest speaker: Michael Lopez, Director of Football Data and Analytics, NFL

Player-tracking data, where the locations of the players and the ball or puck are given several times a second during gameplay, are now available in several professional sports leagues, and for some college sports. Information extracted from this data can augment existing metrics, like the aforementioned team and player ratings, or can be used to address new problems that would be difficult or impossible without the spatiotemporal information available in these new data. In this session, we will discuss:

  • How player tracking data can be used to automatically and quickly identify game events.
  • How these events and other data can be used to create novel player and team statistics that measure new aspects of their performance.
  • Models that estimate expected points throughout the duration of a play, and how they can be used to evaluate positioning and decision-making of players on both offense and defense during those plays.

Roundtable Discussion

Join Sports Analytics Colloquia speakers for a final roundtable discussion in closing. Registration is not necessary for the final roundable; any program participants who are registered for at least one session may join the roundtable.

Program Experts & Guest Speakers

The Carnegie Mellon Sports Analytics Colloquia program is led by Brian Macdonald, Special Faculty in Sports Analytics in the Department of Statistics and Data Science at the Dietrich College of Humanities and Social Sciences, and former Director of Sports Analytics in the Statistics and Information Group at ESPN. Each workshop session features special guest speakers from ESPN, the National Football League (NFL), the National Hockey League (NHL), The New York Times, and others.

This program is presented by the Department of Statistics and Data Science at the Dietrich College of Humanities and Social Sciences and Executive Education at the Tepper School of Business, and is part of the Carnegie Mellon Sports Analytics initiative.

brian macdonald

Brian Macdonald
Special Faculty in Sports Analytics
Department of Statistics and Data Science, Dietrich College of Humanities and Social Sciences

Previously, Macdonald was the Director of Sports Analytics in the Statistics and Information Group at ESPN; the Director of Hockey Analytics with the Florida Panthers Hockey Club; and an Associate Professor in the Department of Mathematical Sciences at the U.S. Military Academy West Point. He received a bachelor of science in electrical engineering from Lafayette College, and a master of arts and a Ph.D. in mathematics from Johns Hopkins University. Follow Macdonald on Twitter


Christie Aschwanden

Aschwanden is the author of The New York Times bestseller, "Good to Go: What the Athlete in All of Us Can Learn from the Strange Science of Recovery" and the co-host of "Emerging Form,"  a podcast about the creative process. She’s the former Lead Science Writer at FiveThirtyEight, and currently an ideas columnist at Wired and a fitness columnist at Elemental. She’s also a frequent contributor to The New York Times, Washington Post, and Scientific American. Aschwanden has been a fellow at the Santa Fe Institute and the Carter Center, and has won awards from the AAAS Kavli Science Journalism Awards and National Association of Science Writers, among others. Follow Aschwanden on Twitter and Instagram


Brian J. Burke
Senior Sports Analytics Specialist 

Burke is a Senior Sports Analytics Specialist at ESPN and an early pioneer in football analytics. Prior to joining ESPN, he founded the website Advanced Football Analytics, where he developed the core metrics and models still used throughout the sport. He formerly was a consultant to several NFL teams and was a regular contributor to The New York Times, Washington Post, and NBC Sports, among other outlets. His recent work has been focused on building individual player metrics using tracking data. Follow Burke on Twitter.


Paul Sabin
Senior Sports Analytics Specialist

Sabin is a Senior Sports Analytics Specialist for ESPN. He received a bachelor of science in statistical science and French, as well as a master of science in statistics, from Brigham Young University, and a Ph.D. in statistics from Virginia Tech. At ESPN, he has worked on sports analytics projects in the NBA, NFL, college football, college basketball, and fantasy soccer. Among these, he was primarily responsible for several ESPN proprietary metrics, including the college basketball power index, the Allstate Playoff Predictor, the NBA draft model, the college football PlayStation Player Impact Rating, and ESPN’s fantasy soccer projections. Sabin has written for ESPN and FiveThirtyEight, as well as contributing to ESPN’s business by using statistical methods to inform programming decisions. Follow Sabin on Twitter.



Lauren Poe
Associate Sports Analytics Developer

Poe is a seven-year member of ESPN's Statistics and Information Group and currently is the only female on its industry-leading sports analytics team as an Associate Sports Analytics Developer. At ESPN, she works with the team to create analytical storytelling tools, such as the football and basketball power index ratings used across collegiate and professional sports. The Oklahoma native and University of Oklahoma mathematics graduate started her career by blending her love of sports and numbers as a high school math teacher and coach. Follow Poe on Twitter.


Jack Birch
Professional Scouting Consultant
Winnipeg Jets, NHL

Birch currently is a Professional Scouting Consultant for the Winnipeg Jets in the NHL. Since 1989, he has worked as a scout for many teams in the NHL. Prior to working as a scout, he was the Assistant Coach for the New York Rangers. He currently is an Adjunct Professor in the MBA Sport Management program at Florida Atlantic University's College of Business. He received his Ph.D. from the University of Waterloo and researches ascriptive and performance predictors of attainment of success in professional hockey.


Michael Lopez
Director of Football Data and Analytics

Lopez is the Director of Football Data and Analytics in the NFL, and a lecturer of Statistics and Research Associate at Skidmore College. In the NFL, his work centers on how to use data to enhance and better understand the game of football. Follow Lopez on Twitter.

Executive Education at Carnegie Mellon

Contact a program director at 412-268-2304 or to learn more.