November 28, 2023

NBA's AI and Cloud Tech: Enriching Fan Experiences and Enhancing Team Strategies

Author:

Tim Yocum

EVERY NIGHT, WHEN a game in the National Basketball Association (NBA) ends, the next frontier of on-court strategizing begins.

Key Takeaways

In December 2020, the league partnered with Microsoft Azure to launch NBA CourtOptix, a platform that delivers fans next-generation insights behind every shot, pass, and play. It does this by using machine learning (ML) and artificial intelligence (AI) to analyze players’ on-court movements in the cloud—information that is then used to deliver nuanced breakdowns to fans. Now, the league is working with its basketball teams to integrate data through Azure into their individual tracking resources, helping coaches better understand how their players perform on court.

To do this is no small feat.

Take defense in the NBA. Unlike baseball and football, which have defined plays with a clear number of participants, basketball flows uninterrupted, making it difficult to track individual performance. This is particularly true in defense, in which many team members contribute in small ways to a play. With the NBA CourtOptix platform, however, defense can finally be measured more effectively. A camera system captures players’ movements continuously, generating roughly 10 million data points per game. That information is then analyzed in the cloud through Azure. The resulting analysis is transforming the game experience for both fans and players.

Sarachek continued, "our algorithms calculate a variety of inputs, such as whether a defensive player is denying passing lanes or positioned in between the offensive player and the basket, and creates a score that quantifies how much pressure each defensive player is putting on each offensive player on the court throughout the game. We use a variety of AI and machine-learning techniques on Azure to identify defensive matchup responsibilities throughout the game.”

“People often say, ‘Well, there’s no good way to evaluate a player’s defensive ability,’” says Sydney Sarachek, director, stats technology product development at the NBA. “It’s been a huge priority of ours to change that. By building new algorithms using the player tracking data, we have been able to make progress."

By assessing that info after a game, teams are now able to quantify their defensive play and devise new strategies. “It unlocks a whole new level of insights,” says Sarachek.

Using Cloud Data to Deliver Personalized Data On Demand

When NBA CourtOptix was introduced in 2020, the platform was primarily focused on the fan experience, delivering analysis after games that paired video with previously hard-to-track statistics, like which players get double-teamed more. But now, with the help of Microsoft Azure, the NBA can share advanced stats that enrich journalists’, teams', and employees’ understanding of the game which can transform how the game is experienced.

Every game night, NBA teams receive a cache of data on each game, a detailed breakdown that is changing team strategies. Teams that have signed up to receive data get it after each contest, helping them make adjustments on the fly—all thanks to a seamless backend data flow created by Microsoft Azure developers. As soon as a game ends, Azure Cosmos DB is used to check metadata to ensure the system should process the matchup. Then, Azure Kubernetes Service kicks off various pipelines running on Azure Databricks, which leverages ML and AI to process information (like the aforementioned defensive metrics). After being stored on Azure Data Lake Storage, the data is automatically synced to teams’ Azure Storage Containers using Azure Data Share. This cloud-first approach helps the NBA save money by being able to scale resources up and down as needed, while ensuring data is seamlessly processed and shared with teams.

“With Azure Data Share, we can go into the Azure console and invite a new team or partner to receive the data,” Sarachek says.  “Once they accept the invitation, they receive updated data in their Azure environments without having to build workflows or processes to pull it in themselves."

"Our goal is to introduce tools and processes that enable our teams to spend more time building analyses that help them make game or lineup decisions—and less time worrying about infrastructure and set up.”

- SYDNEY SARACHEK, DIRECTOR, STATS TECHNOLOGY PRODUCT DEVELOPMENT AT THE NBA

When the NBA creates a new metric, it needs to fill in the same type of information for past seasons to get a larger context for the statistic, resulting in billions of data points that need to be crunched. While this process may have previously taken an engineer weeks to backfill the data, the league can scale up its compute power and leverage over 1,000 virtual machines at once running on Azure to finish the job in minutes, resulting in significant time and cost savings and making it easy to iterate.

To accomplish this, the NBA needed a cloud-first approach—and a cloud partner that allowed it to scale up dramatically when needed. “Being able to spin up more compute when we need it during games is crucial,” Sarachek says. “We can leverage Azure’s compute and best-in-class machine-learning capabilities without investing in those same resources 24/7.”

Creating New Real-Time Insights for the Future

The true innovation on the horizon, however, may change basketball forever: real-time data.

As mentioned, teams and fans can currently access Azure analytics after games. But the NBA is working to deliver its data instantly to teams, fans, and partners, including broadcasters. It’s a development that will unlock even better in-game analysis.

“We are very excited about the fan engagement opportunities for real-time player tracking data, and confident that everything we have built up thus far on Azure will help us serve those insights to fans during games,” Sarachek says.

Transform Your Business with Cutting-Edge AI Solutions from Yocum Technology Group

In today's rapidly evolving digital landscape, staying ahead means embracing the power of Artificial Intelligence. At Yocum Technology Group, we specialize in seamlessly integrating AI into your business infrastructure. Leveraging the robust capabilities of Microsoft Azure, we empower your organization to:

  • Enhance Data Analysis: Unlock deeper insights and make data-driven decisions with advanced AI algorithms.
  • Streamline Operations: Automate routine tasks, reduce errors, and boost efficiency with intelligent AI solutions.
  • Innovate and Grow: Stay ahead of the competition by adopting AI-driven strategies that foster innovation.

Join us in pioneering the future of business. Contact Yocum Technology Group today to explore how AI can revolutionize your operations. Get started Today

Tim Yocum

About the author

As a software architect with over two decades of experience, I specialize in web development (MVC, Angular, React, Blazor), Azure development and migrations, and mobile development (Flutter). My extensive background includes roles as a custom development consultant, database and data warehouse designer. Actively involved in the Microsoft user community, I coordinate the Mid Atlantic and Lehigh Valley Azure user groups. I frequently present at regional user groups and webcasts, and, as a U.S. Navy veteran, I am keen on connecting with fellow veterans on LinkedIn.

What Are You Waiting For? Grow Your Business Today!