We communicate by telling stories. In our professional world, we often use data to support our stories. But how do 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 judgment. 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 policymakers, 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 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 have a very polite shrug. A compelling story reveals insight, not just 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 its origin story. 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.

Most of the context isn’t in the data—it’s 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.

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—whatever works. Context is everything. Stakes are everything-er.

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. 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. 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. 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?

Here’s a data-driven report on seasonal variations in fuel efficiency, with outliers removed for clarity:

This report analyzes fuel efficiency trends across seasons using cleaned gas consumption data. Outliers were excluded using the interquartile range (IQR) method to ensure accuracy and reliability.

Average MPG by Season

• Winter: 29.34 MPG
• Spring: 29.83 MPG
• Summer: 30.45 MPG
• Fall: 29.45 MPG

Insights

1. Summer leads in efficiency: With an average of 30.45 MPG, summer months consistently showed the highest fuel efficiency. This is likely due to favorable driving conditions—dry roads, fewer cold starts, and more highway travel during vacations.

2. Spring shows improvement: Averaging 29.83 MPG, spring reflects a transition from winter’s inefficiencies. Warmer temperatures and smoother driving conditions contribute to better performance.

3. Fall remains steady: At 29.45 MPG, fall maintains solid efficiency, though slightly lower than spring and summer. This may be due to cooling temperatures and a return to routine commuting.

4. Winter lags behind: With the lowest average of 29.34 MPG, winter’s cold weather, increased idling, and potentially hazardous road conditions reduce fuel economy.

Conclusion

Seasonal variations in MPG are modest but measurable. The data suggests that environmental factors—temperature, road conditions, and driving patterns—play a significant role in fuel efficiency. For vehicle owners, this highlights the importance of seasonal maintenance and adapting driving habits to optimize fuel usage.

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.