Reflections from Forecasting for Social Good
Data Lab for Social Good, Cardiff Business School, Cardiff University
I lead an active research group with around 12 PhD students, collaborating with:
Looking back, the journey from uncertain lecturer to research leader did not happen through a grand strategy.

A few years ago, I visited a health centre in West Africa.
At first glance, it looked like a logistics problem. But the real issue wasn’t manufacturing or transportation — it was forecasting.
Planners simply didn’t have reliable tools to anticipate demand.

Standing there changed how I thought about research. Until then, I had been doing what many of us do:
But the real problem was not a lack of models.
The real problem was a lack of translation between research and decision-making.

How can forecasting research actually help people make better decisions under uncertainty?
That question eventually led to Forecasting for Social Good.
Solving these problems wasn’t going to happen through one paper, one model, or one project.
It required partnerships with:
And that meant my role had to change.

I realised I was no longer just a researcher. I had to become something else:
What started as a research idea became a programme combining research, collaboration, and training across multiple countries.

Looking back, I wasn’t following research trends. Instead, I was following:
The gap between forecasting research and real decisions fascinated me. The idea that research could empower people to make better decisions excited me.

As researchers we are trained to optimise models. But in real systems:
A slightly simpler model that people actually use can be far more valuable than the perfect model that sits in a paper.

Academic success often moves fast: papers, citations, outputs.
Real-world impact moves differently. It requires:
Some of the most impactful work I’ve done took years before producing a paper — and sometimes it produced no paper at all, but it changed practice.

At first I thought impact would come from better models. But over time I realised:
The most scalable form of impact is people.
If you train analysts and practitioners, they will continue applying those skills long after a project ends.
Sometimes the most powerful thing a researcher can do is build capability.

There is also a less visible side to research leadership. The more successful a programme becomes:
At first this felt like moving away from research. But leadership means enabling knowledge, not just producing it.

Many societal challenges require collaborative research ecosystems.
But academic systems still reward mostly individual outputs.
How do we build academic careers that reward building ecosystems and capacity, not just producing papers?
Moving from researcher to research leader is not about becoming more important.
It’s about making your research less about you — and more about the problem you’re trying to solve.