Good value for audiences – data science in media today

This time we're taking a look at an increasingly important role in media organizations: data scientists!

Data science is often about generating insights and intelligence to improve the running of a business. That’s true in the media industry too.

But increasingly, and especially with the advent of AI, data science in media is about building new capabilities and services that create added value for audiences. One example is accessibility technology for content on a large scale, such as subtitling. And in times of disinformation, filter bubbles, deep fakes and a general erosion of trust, data science with its pursuit of objectivity offers some of the more promising ingredients for antidotes.

That’s why the EBU and the expert communities connecting through it maintain a lively exchange on data science and AI. Best practice is discussed, solutions exchanged. Ethical standards are established. New skills are built and transferred, especially for the powerful new toolbox of AI technologies. (A good place to see the action is the EBU Data Technology Seminar, 12-14 March!)

One of the principal data scientists organizing this work at EBU is Alexandre Rouxel.

Here's what Alex says about the field and how it's changing:

"We see media organizations hiring more data scientists to work on a range of things. But generally, you'll find yourself at the coalface of the digital transformation of that organization.

The field is constantly evolving. We need a diverse skillset that encompasses statistics, software development, cloud architecture, and data modeling. To thrive in this landscape, data scientists are usually multi-faceted profiles, blending analytical rigor with creative intuition.

We generate capabilities and insights. Data science can also help media organizations to create unique services that help the public to form opinions.

And the EBU is the place where data scientists from across the membership come to share this knowledge, their ideas, to learn from one another and collaborate on common solutions. It’s creative, it's exciting!

Science and creativity? How does that go together?

You might say that we work to replace intuition or gut feeling. But our job is not just about crunching numbers; often, our job is to transform raw data into valuable insights and enlightening narratives. We extract engaging stories from messy data by unearthing interesting insights.

In a media organization, data scientists are also tasked with understanding, structuring, and adding value to the vast amounts of data contained in the content that is produced and how it is being consumed. By analyzing audience engagement metrics, we can help to optimize content priorities and to ensure that every piece of content resonates with our target audience.

And often, we collaborate with other developers or engineers to to create solutions that add to the capabilities of the company.

The big one: AI

Unsurprisingly, data scientists were among the earliest creators and adopters of AI. We use it extensively now. Not the large language models and generative AI that aim to do everything: those options target end-users and are often too general, expensive, energy intensive and slow to use at scale.

Instead, we choose the right tool for the job. That's often a smaller, more targeted model that we’ve set up to efficiently do very specific things. The technology is now at a point where we can develop production-ready specialized AI tools in a fraction of the time it once took.

Maximising the upside

Of course our job is also about knowing the landscape and helping our organizations to navigate it. We help to balance the potential benefits of cutting-edge technology with considerations of cost, feasibility, and of course ethics.

And because the pace of AI innovation is accelerating, data scientists have to keep up with it too. But that’s exactly what we must do in public service media – in order to help ensure that there’s a significant upside in the AI revolution for the general public. That, ultimately, is our mission.