Improving at a sport requires good training practices and tracking of progress. At a semi-professional level, players do not have access to high-quality equipment for data capturing nor dedicated analysts picking out areas for improvement.

At the moment, the improvements can only be suggested based on visual references and observations taken by the coaching staff, who have very limited capacity.

There are 20-30 cameras in any given professional game for data generation and sports tracking

We wanted to build something low cost that is highly accessible through technology that is currently available and widespread. 1 in 6 Australians wears a smart device (e.g. sportwatch), and 1 in 6 Australians participate in a team sport. These facts motivated us to utilise wearable smart devices and machine learning technology to improve the team sport experience. Ultimately, we want to remove the limitation of access to equipment and high costs, which can be a barrier of entry for smaller-budget sport teams who want to improve.

What it does

There are three main steps in our application.

  • Capture sporting data (acceleration, field position, heart-rate, etc...)
  • Process data in the cloud through smaller modules to generate insights
  • Present and show actionable insights for given players in a user-friendly user interface

A private school focuses on sporting excellence as its distinguishing factor to attract top-performing students. Their sports department has moderately adequate funding to buy equipment to better the facilities the school can provide.

If data analysis is to be undertaken, it requires specialist analysts to be hired alongside expensive equipment to generate data, costing in excess of $100,000 (20 cameras + 2 data-analyst and camera-person) to provide the same levels of data capturing and tracking in equipment and personnel.

Each professional AFL game costs around $300,000/hr to capture sports data for the players.

We simplify data collection using a smartwatch and provide a service to run analytics to get detailed insights using computational and artificial learning methodologies in the data analysis section.

How we built it

Smartwatch is used to capture raw data which is stored in a remote (Firebase RTDB) for future retrieval for processing. It's stored in realtime to allow real-time analysis too if required.

The data from there is fed into microservice (components) to analyse the data which outputs the insights.

The processing engine is written using Springboot and deployed on AWS cloud using containerisation. (Docker)

The Front-end android application is built using Android SDK and other utility libraries listed in the dependencies of the project file.

Challenges we ran into

When we started with the project, we did not have knowledge about passing the collected data from the smartwatch to the application. However, with some researching and trying stuff out, we were able to figure out how to implement the data transfer system and storing it in the cloud.

Accomplishments that we're proud of

  • Making our first smartwatch integrated application that collects and provides data instantaneously.
  • Made use of EchoAR which creates a virtual field that can help the coach in guiding the players better.
  • Actually finishing a reasonably workable prototype
  • Real-time updates and analysis

What we learned

Throughout the development of this project, we were able to learn a lot of things. Especially, we learned the implementation of EchoAr and Twilio in the application.

Time management is really important. Architecting a cloud application is quite vital in terms of the data flow to make it future proof as well.

What's next for

We hope to expand it by incorporating more sports and features. Like focusing on sports like cricket, boxing where the watch can collect data like swing motion, punch speed, and other necessary attributes required for the individual sport. In the future, we hope to reach a point where it will be marketable and can be used by any sportsperson.

Hiring based or Subscription-based analysis can be started as a service to bring in revenue and hopefully add a few more sports along the way.

Final notes

We all teammates did not know each other but really enjoyed this experience 😊

Costs of video capturing

Built With

Share this project: