Daniela Stevenson

Let’s start with you. Who are you, and what lens do you bring to understanding people online?

With a background in web development, I bring a builder’s mindset to social intelligence, using APIs and automation to uncover insights that standard dashboards miss. 

More recently, I’ve been experimenting with prompt engineering to enhance every stage of social data analysis, from generating smarter queries to automating topic clustering and sentiment evaluation. Over the past year, these experiments have significantly improved in both accuracy and relevancy, allowing me to work faster while keeping nuance intact.

‍

What’s a working theory you have right now about how people behave online?

While polarisation gets a lot of attention, what often matters more is where conversations are happening, the same community can sound radically different on TikTok versus X or Facebook. Looking at social data in aggregate can flatten those nuances and miss what’s actually driving behaviour.

That’s why I think it’s critical to build a platform by platform picture rather than assuming insights translate cleanly across channels. It also means ensuring your tool stack allows you to actually listen where your audience is most active.

Some of the most honest conversations often happen in closed platforms, like Discord servers, where superfans hang out. Keeping an eye on these spaces can reveal what people really care about and not just what performs publicly.

‍

What’s an insight you surfaced that you still think about? What one stuck with you?

One insight that still sticks with me came from a project I did purely for fun. To help demonstrate what our social tools can uncover, I did a light-hearted study into how many pirate terms are still hiding in plain sight in everyday language.

Armed with graphs, verbatim examples, and a bit of curiosity, I was able to show how phrases like “all hands on deck” and “learning the ropes” are still very much part of modern conversation, even if most people never stop to think about where they came from.

What stayed with me was how effective it was as a teaching moment. By having a bit of fun and leaning into the theme, the value of social listening became much clearer and hopefully left people feeling like they’d uncovered some hidden treasure rather than sitting through another demo.

‍

What’s the weirdest rabbit hole your work has ever sent you down? And what did it teach you?

One of the things I enjoy most about agency life is the variety. I get to work across very different projects, from functional drinks and toys to major streaming shows, and switch between trends work, conversation analysis, sentiment, and even crisis support depending on what’s needed.

The weirdest rabbit hole I’ve gone down recently came from a movie release where I was tasked with finding potential influencers. That search unexpectedly dropped me into the world of terrarium artists, creators who build and document incredibly detailed miniature ecosystems, and the highly engaged communities that follow them.

It was fascinating to see both sides, the care and craft that goes into these tiny worlds, and the audiences who are deeply invested in them. On a broader level, it reinforced the idea that there really is a niche community for everything. On a practical level, it also taught me not to underestimate how much time it takes to properly vet creators, reach out, and secure them. 

‍

What skills or mindsets do you think the next generation of analysts will need?

As AI-powered tools become more common, it’s going to be much easier for anyone to access and summarise data at scale, including social data. That’s a really positive shift, but it also changes where the real skill sits.

I think the next generation of analysts will need a strong understanding of the inputs, not just the outputs. Writing a good social listening query that returns results at volume and with relevance still takes judgment, context, and experience. If you don’t understand how the data is being collected, it’s very easy to misread what comes out the other side, especially when AI is doing a lot of the work for you.

Beyond technical skills, curiosity and critical thinking will matter even more. Analysts will need to question summaries, spot bias, and sense-check insights rather than taking them at face value. In a world where data is easier to digest, the real differentiator will be understanding what should be included, what’s missing, and why.

‍

What’s a niche community, account, or corner of the internet you’re watching right now? And why?

As a Vinted user myself, I’ve become invested in the platform’s unwritten rules and social norms. There’s constant debate over what’s acceptable behaviour, from whether it’s okay to chase buyers to collect a parcel before the clock runs out, to the strategy of deliberately overpricing items in anticipation of the inevitable discount requests.

Alongside the more serious conversations about scams and trust, there’s a steady stream of humour: chaotic packaging choices, scarves modelled on a Harry Styles cardboard cutout, or even bus stops listed for sale.

It’s a fascinating micro-culture where commerce, etiquette, and meme-making collide and shows that even the most functional platforms develop rich, self-policed social systems worth paying attention to.

‍

Last non-work thing you read that shaped your thinking?

Babel by R.F. Kuang (which I’m currently reading with my book club).

There’s a part in the book that debates whether translators should be faithful to the author or to the reader, and it really stuck with me. Translation isn’t neutral, choices about framing, emphasis, and omission subtly shape meaning.

It’s made me think a lot about our role as translators of social data. When we turn messy, contextual human expression into insights, we’re making similar decisions: what gets foregrounded, what gets smoothed over, and who we’re ultimately translating for. What’s lost in that process? Whose perspective are we privileging?

It’s been a useful reminder that interpretation always carries power, and that being intentional about audience and bias is just as important as technical accuracy.

‍

‍