When Google wanted to demonstrate the power of their Cloud Platform – they turned to Sportradar.
Using Sportradar’s raw 2016 game logs, Google Cloud Platform created a public set of denormalized baseball data within its BigQuery analytics warehouse. Data buffs were invited to use BigQuery to parse and query the Chicago Cubs, Cleveland Indians and all the World Series action down to the pitch level to see if they could score analytics magic.
Google Cloud Developer Advocate, Eric Schmidt, detailed how GCP Inside Baseball – aka Project Uecker – was built and chronicled his insights throughout the World Series.
“A key goal of Project Uecker was to create highly accurate analysis in near real time. The fidelity and speed of the Sportradar API enabled us to create complex, long running workflows while still being able to implement ad hoc, exploratory analysis. We were able to gain meaningful insight in tandem with the game flow, bringing users closer to the micro moments of the game.” — Eric Schmidt, Lead Developer Advocate at Google
A good example of how Eric utilized Sportradar’s data to inject interesting insight in real-time occurred in Game 7 when Chicago Cubs pitcher Jon Lester threw a wild pitch that resulted in two runs scoring. Eric was able to quickly calculate and share that there were 716,708 pitches thrown in the 2016 season with only two other instances in which a wild pitch led to two runs being scored. That equates to a 0.003% chance of that specific play would happen.
“We were blown away by the insights that Eric and the Google Cloud Platform were able to uncover in seconds. Project Uecker was a testament to the power of big data and an awesome way to leverage the enormous interest in a historic World Series,” — Tom Masterman, Chief Revenue Officer at Sportradar
The analysis can from Project Uecker can be read on Google Cloud’s Medium page.