Tune into Data

Tune into Data - AI and Azure: how a data scientist analyses 45 billion spins

Posted by Future Anthem on December 20 2021

At Future Anthem, we’re always trying to create and share insights into player behaviour, to deliver actionable recommendations.

In our final Tune Into Data of 2021, the session was a chance to lift the lid on how our data scientists operate – to find out about the tools, processes and technology they utilise – as we hear from Future Anthem’s Ed Mitchell, Jorge Brasil, and Younes Sandi.

We’re spoilt – we have no legacy tech; just completely green fields,” says Ed. A far cry from some of his earlier experiences at big-name PLCs where a wrong move in the server room on an antiquated Java machine could take down the entire infrastructure.

This allows us to identify and use state of the art tech and analyse a mind-blowing amount of data – such as billions and billions of spins – in real-time.

The software is now taken for granted after speed-of-light developments in recent years and the possibilities it enables are virtually endless.

Why Microsoft Azure?

Future Anthem runs on Microsoft Azure with the team working on data bricks. Watch as Ed explains the benefits, including from a security perspective:

 

He then goes on to discuss Future Anthem’s real-time data infrastructure. One of our greatest assets, it allows us to pass in-game information into our models in milliseconds and output an instant result, while a player is still playing.

This significantly boosts amplification opportunities and player safety, as Ed details:

 

For companies thinking of creating a cloud-based solution, Ed shares some hints and tips that the Future Anthem team have learned:

 

Creating top-tier RG models

Jorge’s work on responsible gaming (RG) has made a vital contribution to a key product offering, Anthem Safer Play, which proactively detects players at risk based on in-game behaviours.

As a focus area for Jorge, he has brought a new way of thinking, which has helped to overcome some of the difficulties found with supervised learning techniques that had previously been the traditional approach to RG modelling.

Our approach incorporates ‘clustering’. Watch as Jorge gives an overview of how it works:

 

Putting theory into practice with Gamesys

A high-profile application of Anthem Safer Play has been with Gamesys, where modelling of 36 million gaming sessions found no statistically significant correlation between player markers of harm and three structural game characteristics: Volatility; Return to Player (RTP); Hit Rate.

As Younes says this was groundbreaking research: “None of the previous papers we encountered have done research in this area. None had investigated the impact of different slot games on players.

He describes how the team did it and what we found:

 

If you’re interested in finding out more about our game data science platform or booking a demo, please get in touch.

You can also connect with Ed, Jorge and Younes on LinkedIn:

AI and Azure: how a data scientist analyses 45 billion spins: the full webinar is available to watch here.

 

Join us again in 2022 when Tune Into Data will return with more events that are designed to educate, entertain and inspire.

 

Gamesys Download CTA