We communicate by telling stories. In our professional world, we often use data to support our stories. But how can we tell compelling stories with data? This post explores, very briefly, the art and science of data storytelling, providing insights and techniques to help you craft narratives that resonate with your audience.

The Elements of a Good Data Story

There are no hard rules for good writing—just a few trusty guidelines and a lot of judgement. Consider these. Use them wisely. Break them with flair.

Who, What, and Why

Audience

Every good story begins with knowing your audience. Who are they? What keeps them up at night (besides grading)? What do they already know? Or think they know?

The story you tell your faculty advisor will not be the same one you share with policy makers, journalists, or peers. This isn’t about changing your findings; it’s about spotlighting different parts of the elephant depending on who’s asking. There’s no single “right” story, only the right story for the this crowd, at this time, for that purpose. Sometimes, maybe two stories.

Transformation

Great stories are about change. What did we learn? What problem did we solve? What new questions popped up like weeds?

If your research confirms what everyone already suspected (but hadn’t bothered to prove), you don’t yet have a story. You are confronted with so what? A compelling story reveals an insight or, not just provide validation.

Purpose

A good data story has a mission. What are you trying to do? Convince, inspire, provoke, confuse and complicate? Confusion is a valid purpose. It engenders deeper understanding and sweeps away simplistic beliefs.

What emotions do you want to stir? What actions do you want your audience to take? (e.g., give you money, cite your work, name their first born after you?).

This purpose does not have to be earth shattering, paradigm shifting. It can be as simple as “I want my audience to understand that X is more complicated than they thought.” Or “I want my audience to see that Y is a problem worth pondering about.”

Should your audience be moved? Impressed? Outraged? A story without a purpose is just a long walk with no destination. That’s fine for Kerouac or Knausgaard, but not for you.

The Process

Data Exploration

Spend time with your data. Take it out for coffee. Ask it about where it began? How did it evolve and mature? What is its role in the world? Understand how it was collected, what it claims to measure, and what it’s actually measuring. Who is interested in collecting this data and why? Most of the context isn’t in the data. It is in the metadata, the instruments, the people wielding those instruments, the time spans, the units, the scales, and the deeply human messiness behind it all. Think hard about what is missing, what is biased, and what is just plain weird.

All data are simplified, abstracted, and mangled representation of a reality. And there are many realities, if I want to be post-modern about it. Treat data with curiosity and a healthy dose of scepticism.

Depth

A good story doesn’t try to say everything. It goes deep, not wide. Anchor your story with a thesis. However, you won’t get to a thesis in the beginning. You might want to start with 4–5 core findings that matter. These might include:

• Surprising results that defied expectations
• Patterns that emerged across analyses
• Anomalies that begged for explanation
• Methodological wizardry you invented (In this class, we don’t care about this as much. Maybe in your phd.)
• Theoretical frameworks you tested or refined
• Real-world applications that actually work

For each finding, ask:

• What did we believe before?
• What changed?
• What new questions arose?
• What became possible?

If nothing changed, toss it. Ruthlessly. Your story deserves better. Your audience deserves better.

Eventually, settle on 1–2 findings that carry the most weight. These are your story’s spine. They inform your thesis.

The Presentation

Your audience needs to know what’s at stake. Set the stage at the beginning. Make them care. Use a motivating example, a neglected concept, a spicy newspaper headline. Context is everything. Regardless of your earth shattering findings, if no one cares, the story does not matter.

Structure

Every story has a beginning, middle, and end. The middle is where your methods and results live - and let’s be honest, only your committee cares about this. Much of your graduate training is about this middle.

However, the important things are the other two. The beginning hooks your audience. The end leaves them thinking, feeling, and maybe tweeting about it.

Exhibits

Just because you can make a pretty chart doesn’t mean you should. Just because you compute an average to 6 degree precision, does not mean you should. Every exhibit—table, figure, map, interpretive dance should serve a purpose. It should clarify, not confuse¹. Label it clearly. Integrate it into your narrative.

Remember, exhibits are not neutral. They are not faithful representations of your data. They are evidence to support your argument. As such, they are heavily biased towards that purpose. Lean into those biases. Use them wisely. Use them effectively. Think about what you need to foreground and what you need to background. Think about what you need to highlight and what you need to suppress. Think about what context you need to provide to make them legible. When you come across these exhibits in other’s stories, ask what biases they encode. What’s being highlighted? What’s being hidden?

Language

Speak your audience’s language. If you’re presenting to TikTok, bring memes and dances. If you’re writing for academics, dust off your jargon. Check out these tips, I scribbled down back couple of decades ago. Some of them might still be relevant.

Editing, Presenting, Reworking

A good story is forged in the fires of revision. It evolves. It gets feedback. It gets worse before it gets better. Writing isn’t what comes after analysis. It is integral to it. Through writing, we clarify, wander, connect, and often realise we were wrong, all along. Edit ruthlessly. Kill your darlings. Rework what doesn’t work. Repeat.

An Exercise

Because I am quirky (a polite way of putting at it), I have been writing down information, whenever gasoline is added to my car.

Take about 30 minutes to explore this data and paint a picture. Use whatever tools, software as you see fit.

Here is what Microsoft Copilot ’thought’ the story is. Can you do better?

From 2020 through 2022, the vehicle behaved with remarkable consistency. Most fill-ups were about 10 to 12 gallons, and most occurred after roughly 300 to 350 miles, implying fuel economy and driving patterns that barely changed over thousands of miles. The car accumulated approximately 33,000 miles during this period, moving from about 131,000 miles in October 2020 to 164,000 miles by the end of 2022, which works out to roughly 16,000 miles per year. Then 2023 arrives and the pattern breaks. In just the first seven months of the year, the odometer jumps from 164,070 to 175,873 miles, nearly 12,000 miles, a pace about 30% higher than the preceding years. The change is not subtle. January alone contains an unusual cluster of fuel purchases, and March 2023 shows repeated refueling over consecutive days as the vehicle moves from 169,764 to 172,282 miles in less than a month. The data suggest a year of exceptional mobility, when the car was being used far more intensively than before.

But the most interesting part of the story is that the surge does not last. After reaching 181,728 miles in May 2024, mileage growth slows dramatically. Over the next eight months, the vehicle adds only about 3,700 miles, a fraction of its earlier pace. Long intervals appear between purchases, including a gap of more than two months during the summer of 2024. At the same time, small purchases become increasingly common: tanks of 5 to 8 gallons begin to appear where full 11-gallon fill-ups once dominated. By the final chapter of the dataset, the car is approaching 198,000 miles, but the real story is not the distance traveled. It is the shift in behavior. The data reveal three distinct regimes: a stable commuting phase through 2022, an unusually high-mobility phase in 2023, and a low-utilization phase beginning in 2024. Without recording a single meeting, move, job change, or life event, the gasoline log captures a major change in how the vehicle, and likely its owner, used time and space.

Conclusion

Storytelling is a skill. It takes practice, reflection, feedback, and a lot of reading. My hope is that this course helps you build that skill. And that you’ll tell stories that matter—stories that change minds, spark action, and maybe, get you that grant.