PitcheroGPS Revolutionizes Sports Performance Tracking with SingleStore on AWS

Based in Leeds, U.K., Pitchero has been a pioneer in providing technology solutions to sports clubs and teams for more than a decade. With its latest venture, PitcheroGPS, this fast-growing company introduced a groundbreaking GPS tracker designed to monitor athletes' performance during outdoor team sports such as soccer, rugby and field hockey. The player tracker is already approved by FIFA and World Rugby for use globally.

“Our core business has been around for about 15 years providing technology that helps sports clubs and teams spend more time playing and less time on administrative work,” said Jon Milsom, Co-Founder & CTO, Pitchero. In July 2023 the company launched PitcheroGPS, its first hardware product: a GPS tracker worn by soccer, rugby and field hockey players (or any outdoor team sport) to monitor their performance during the game and allow them to track their progress or compare against teammates. With a vision to democratize access to advanced sports data analytics, PitcheroGPS aims to empower athletes at all levels.

Challenges + goals

Pitchero faced the challenge of efficiently handling vast amounts of time series and geolocation data generated by hardware trackers. This generates hundreds of millions of rows with numerical and geographical data that need to be aggregated and parsed quickly. Traditional relational databases like MySQL were not up to the task, leading the team to seek a more robust solution capable of processing hundreds of millions of rows swiftly. “After 15 years of experience with the product, the team knew MySQL would not cope well with this type of workload,” said Milsom.

Technology requirements 

"Fast query times across large datasets were imperative for us, considering the sheer volume of data generated per session," said Milsom. “Our hardware tracker samples at 10Hz (10 times per second), so a game or training session lasting 90-180 minutes can provide 50,000-100,000 rows of data, per individual, per session.”

Other advantages he and the team were looking for included:

• Fast query times across large datasets (~ hundreds of millions of rows) 
• Fast computation across the time domain (TIME_BUCKET()) 
• Ability to perform simple mathematical instructions quickly (e.g.. AVG(), MAX())
• Ability to run in the cloud

Why SingleStore

Pitchero knew it needed to bench MySQL and put a high-performing data platform in the game. Milsom had read Why Fathom Analytics ditched MySQL, Redis, and DynamoDB to power its website analytics platform. We interviewed Fathom Analytics Co-Founder Jack Ellis for that one, and appearing at the bottom of that page are several blog posts Ellis authored describing how he and the team built the world’s fastest website analytics with SingleStore. “We were able to make a proof of concept using the SingleStore Docker images and were able to assess the raw speed with a large number of rows of data,” said Milsom, “but to be honest, we were already very confident after reading the Fathom Analytics case study.”

Pitchero also considered MongoDB Atlas and an AWS version of MongoDB, Amazon DocumentDB. “We also considered MongoDB (we use MongoDB Atlas for an analytics workload), but with most of our development team reading/writing standard SQL fluently, SingleStore provided all of the speed with a comfortable query language (SQL),” added Milsom. “We only needed to learn SingleStore-specific commands such as PIPELINE, rather than the MongoDB query syntax. We gained a relatable syntax for our existing development team with no huge cognitive leap.”

SingleStore emerged as the preferred solution due to its remarkable speed, scalability and compatibility with standard SQL. The platform's ability to handle complex queries across massive datasets in milliseconds, coupled with its ease of use for the development team, made it the perfect fit for Pitchero's requirements.

Solution

Pitchero deployed SingleStore Helios, leveraging  advanced platform features including:

• Time Series Aggregation functions
• Pipelines
• Unlimited (Bottomless) Storage

The team seamlessly integrated SingleStore with its existing technology stack, including Laravel Vapor for AWS Lambda and Golang for data processing. “We run one production cluster and one staging cluster with segmented workspaces and workloads,” said Milsom. “We can switch off the staging cluster to save money when we aren’t using it.”

Data from the hardware trackers is uploaded to Amazon S3 via the PitcheroGPS mobile app, and SingleStore Pipelines 

ingest the data into SingleStore for analysis. As this story goes live, Pitchero is managing over 900 million cumulative rows of session data in SingleStore, and continues to analyze new and existing data to deliver insights to users.

Outcomes 

Since implementing SingleStore, Pitchero has experienced significant improvements in data freshness, query latency and overall performance. PitcheroGPS is now ingesting up to 140,000 rows per second and executes complex queries on nearly a billion rows in milliseconds, powering real-time analytics. “Our customers get super-fast processing of their data so they can see their sessions almost immediately after upload,” said Milsom.

This enhanced capability empowers athletes with immediate insights into their performance, while also positioning Pitchero for substantial growth and expansion into new markets. “We will scale out our sales function in 2024 and consolidate legacy workloads and applications into the application that uses SingleStore,” concluded Milsom. “This means many more customers, many more individual tracked sessions, and much much more data.”

---

We hope you enjoyed this story about how Pitchero is deploying SingleStore’s high-performance data solution to help athletes achieve top performance. Learn how another team of visionary innovators, the team at Fathom Analytics, is gaining operational and competitive advantages with SingleStore.

SingleStore helps companies compete and win across every vertical. Learn more →