Big data helping Aussies to Olympic fast lane

TECHNOLOGY
Cutting-edge data analytics was key to helping the Australian swim team bring home the most successful Olympic medal haul in history last year, says Swimming Australia’s Jess Corones.

First published on Westpac Wire

Sports scientist Jess Corones working with Olympic swimmer Emma McKeon. (Getty)

Jess Corones is the first to admit she’s not a “tech person”, but the sports scientist is certain the cutting-edge data analytics technology she’s spearheaded at Swimming Australia was critical in helping the national swim team last year bring home the most successful Olympic medal haul in its history.

“100 per cent, it helped us go into the Tokyo Olympics as probably the best prepared country,” says Corones, who’s worked with Australia’s elite swimming coaches and athletes for the past decade. 

“Being able to siphon our data for our coaches meant they knew much more than any previous Olympics about the situation they were going into, they understood the competition, so they knew what to train their swimmers for.”

The key, she says, was a transformation in the way Swimming Australia leverages data, developed in the few years leading up to the Olympics, including standing up the first race analysis system in the world to use artificial intelligence and machine learning, named Sparta 2.

"How we work has massively evolved, and part of that evolution has been due to the technology partners we’ve brought on,” Corones told Westpac Wire, ahead of a keynote talk at Westpac’s recent TechX, an annual internal technology convention.

Listen to Jess Corones in conversation with AWS’s Justin Lennon, at Westpac’s TechX: here

A turning point was the partnership struck in 2018 between Swimming Australia and cloud computing giant Amazon Web Services, says Corones. The first joint project off the blocks was the creation of a “data lake” which transformed the way her team processed the information core to their work with coaches and athletes. 
​
Like most sports scientists around the world, Corones had historically collected data manually, which meant when coaches asked her questions on the pool-deck – like ‘What’s the fastest swimming turn time in the world?’, or ‘When a swimmer has a stroke rate of x what will their velocity be?’ – her response was only as fast as her ability to go in search of and analyse the data.  

“I used to spend a lot of time just searching on web pages, for things like international competition data, pulling numbers off and putting them into my spreadsheets,” says Corones. 

“It was really siloed, prone to human error, my training data wouldn't speak to my competition data. I had to draw parallels, but the data couldn't do that itself.”

Jess Corones, Swimming Australia’s Performance Solutions Manager. (Emma Foster)

​Fast forward to today, she has real-time data at her fingertips, drawn from Swimming Australia’s cloud-based data lake, a centralised repository which pulls and analyses statistics from a wide range of sources.

Aside from pulling in competition results from the likes of international swimming federation FINA, and training data from various athlete management systems and wearable technologies, Corones says most useful information in the lake is the granular data captured by Swimming Australia’s world-leading race analysis technology, Sparta 2. 

In development for around five years before its release in 2020, Sparta 2 draws data from championship swim meets around the world where a specially designed 4K camera is used to film each race. Image recognition technology with AI and machine learning is then used on the video footage to analyse every swimmer: their start times, turn times, velocities, distance per stroke, distance travelled under water, number of kick counts, number of breath counts, breath patterns and so on. 

“It's definitely been a really big organisational shift for us,” she says. 

“The coaches and the rest of the teams are starting to get an understanding of the power of cloud and of data – that if we can get it in real time, aggregate it and it’s not siloed, it’s speeding up our processes. 

“Most importantly, it helps us to develop individual race plans for our athletes so that they can get the optimal performance outcome at the end.” 

Corones describes as another “huge breakthrough” the Relay App developed in partnership with AWS. Hooked into the data lake and using predictive analysis and machine learning, the Relay App helps coaches make strategic decisions about swimmer selections for relays for the highest probability of a podium finish – a project Corones says was held tightly under wraps until the Tokyo Olympics. 

“The way the relays work is you swim a heat in the morning and then you can make changes to the team to swim the finals that night. So once the heats are done, we have to make really quick decisions around who we're going to swim in the finals,” Corones says. 

“We always go in with a strategy, but we've got to be able to pivot pretty quickly depending on what happens on the day. The relay app allows the coaches to test their strategies and play out different scenarios, using its predictive capabilities, to ensure we're really making the best decisions.”

AWS’s Justin Lennon interviewing Jess Corones at Westpac’s TechX internal technology convention.

​Corones says the app shone in the highly tactical mixed medley relay at the Tokyo Olympics, a race in which the Australian team was not expected to medal.  

“We made a change at a much later stage than we normally would have based on the data, and they got up on the podium, which was so amazing to see,” she says.

“In fact, we were the highest medal winning relay team at the Olympics. We medalled in six out of the seven relays, which I think only has been beaten once before. It was a huge, huge feat.” 

A more recent project with AWS has been the development of an online portal for coaches, called Lane Four, which Corones expects to be rolled out next month.  She says it’s been designed to make data more accessible for coaches – of which there are around 40 at the senior high-performance level and a further 250 contributing to the Olympic swimming teams.    

“Coaches are not office people; their environment is a pool deck,” she says. 

“They're really savvy with an iPad because they can have that on the pool deck. So now while they’re in a coaching session, they can pull up what they need to know, in real-time, in a much easier way on their iPad and that’s going to shift their training environment to enable athletes to be better trained to compete.”  

Corones is keen to stress that although data helps performance, it’s ultimately about the humans delivering it. 

“Swimming's becoming more and more competitive – you have to train hard, but everyone's training hard for the Olympics, so then you have to get smarter and make sure that your training is the right training. That’s where the data is really, really valuable for us,” she says.

“But all of our coaches and athletes are just people, and I will always put the person first, understand them and what works for them. That's the most important thing.”