Deploying AI Models at the Edge

AWS hopes to score from the edge with Bundesliga Match Facts

AWS hopes to score from the edge with Bundesliga Match Facts

By Daniel Beazer, Contributing Analyst at Structure Research.

AWS put out an interesting case study with a sports customer, Bundesliga, the body that oversees the German premier football league. Amazon confirmed the customer win last spring and Bundesliga is now using AWS to provide real-time statistics and insights, a service it calls Match Facts.


Sports broadcasting is a natural fit for cloud-enabled, locally delivered machine learning and analytics, and organizations like the Bundesliga look to live interaction to drive greater fan engagement and consequently revenues. So, for example, fans can see data points like average player positions, attacking zones, and most pressed player in real-time.

How data is gathered

The most advanced feature the service offers is xGoals or expected goals. Bundesliga collects 3.6m data points during each match, gathered from 20 cameras in each stadium. The machine learning team uses an AWS SageMaker algorithm to analyze over 40k historical shots at goal gathered since 2017, and that machine learning model runs over the freshly gathered data, producing analysis and insights. For example, xGoals will scrutinize a goal and give a scoring probability, 77% for a penalty, or 7% for a remarkable strike from the halfway line.

Analyst Take

Live sports are emerging as an early edge computing and machine learning use case, with the emphasis here being on analyzing data that is generated at the edge. Bundesliga is the latest addition to an impressive list of sporting AWS customers such as NFL, Formula One, and the PGA. But in a way, this latest case study shows the limitations of these tools.

While the machine learning ingest, serverless architecture and CI/CD platform are impressive, do we need an AI/ML platform to tell us that its much harder to score from the halfway line than from the penalty spot? Part of the reason football is so popular around the globe is the sport’s simplicity and the small number of rules. Games can easily end with one or no goals at all, quite the opposite approach from that taken by baseball or cricket where statistics are the lifeblood of the sport. But a game as simple as football does not easily succumb to numerical analysis. Perhaps the availability of these tools in a live format will change that, and it will be interesting to see if they do. Of course, one could also say that the live statistics and insights add an element of entertainment that enhances the overall viewing experience.

About the author

Daniel Beazer is a contributing analyst at Structure Research, an independent research firm focused on the cloud, hosting and data center infrastructure service provider markets.

DISCLAIMER: The views expressed in this contributed post are that of the author, and don’t necessarily reflect the views of Contact us if you want to contribute a guest post. 

Article Topics

 |   |   |   |   |   |   | 


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Featured Edge Computing Company

Edge Ecosystem Videos

Deploying AI Models at the Edge


Latest News