![guy jumping to reach basketball hoop](https://www.siliconrepublic.com/wp-content/uploads/2015/02/shutterstock_432954493-718x523.jpg)
Image: bbernard/Shutterstock
TechWatch editor Emily McDaid sat down with MVPx CEO Justin Thompson to learn more about the stats behind the game of basketball.
Sports performance data has experienced a huge market surge, with the industry predicted to be worth $4.7bn by 2021.
American sports broadcaster ESPN even ranks teams by their grasp of data analytics.
I recently sat down with local entrepreneur Justin Thompson, who’s entering this market with a very American sport: basketball.
What is MVPx?
It gives basketball players the ability to statistically identify weaknesses in their game. It uses computer vision to track their locations on the ball and uses machine learning to decide whether the shot has gone in or not.
What hardware do they need?
Just an Android phone and a tripod.
What’s the challenge in building this?
Getting it to run on a mobile device with limited hardware. For tracking, speed is important, so I need a certain frame rate to get reliable tracking.
I needed to optimise the algorithms to make them run as quickly as possible.
How do you go about building something like this?
The underlying technology is an image-processing library called OpenCV. It’s a very mature, open source framework, very stable and powerful.
I use that, plus my own tracking that I’ve developed on top of it.
Why basketball?
I play all sports but I don’t play basketball much. I was living in Berlin and working in Betahaus (co-working space) on machine learning. It was hard to learn it without applying it. One of my friends was telling me about ShotTracker, a physical bit of hardware that tracks your shots for you. It was expensive and didn’t work very well. It seemed to me that you could do it much better with a camera, so I started working on it as a side project.
Then, the first prototype was way better than I expected. I came back to Belfast and decided to give it a go and set up a company. Basketball is just the first application – it’s a small court with limited players and the rules are clear-cut. But it will eventually work for other sports because the tracking is generic to all sports.
What info does it give the player?
It tracks the players’ speeds, positions, heat maps, and you can see where they’ve taken the shot from, the arc, whether it’s gone in – and it clips out video that the player can review.
What is the golden window?
It’s the angle of release and angle of entry into net. The app measures all your angles. There’s a golden window, where your shot will most likely go in. Training is all about repeatability and getting the golden window.
Can it handle all the different types of shots? Free throws, layups etc?
As long as it can see the net, it can handle it.
Have you trialled it with any teams?
It’s in beta at the minute so I’ll be looking to get a few teams on board, local teams that I can talk to.
Are you developing it all yourself?
Yes, which means I can move quickly. I’m a software engineer by trade, and I previously worked for Aepona and Cloudsoft.
What are your aspirations for MVPx?
I want to take it to other sports. My main aspiration is to bring the price down for people who can’t afford equipment.
Potentially in the next nine months, I’ll look for investors. I’m in a good space now, it’s just me and I can get everything done for now. Funding would be useful once I can’t.
What’s your monetisation plan?
There will be a subscription model for individual players and a team version where coaches can collaboratively look at everyone’s stats.
By Emily McDaid, editor, TechWatch
A version of this article originally appeared on TechWatch
MVPx is a finalist in the annual Invent competition run by Connect at Catalyst Inc, aiming to showcase the best and brightest innovators that Northern Ireland has to offer. Invent 2018 will take place on Thursday 11 October in Belfast, where 12 finalists will battle it out for a £33,000 prize fund.