Everyday, 2.5 quintillion bytes of data are created. That’s a lot of individual bits of information (that’s 2.5 with seven zeros! after it). But bits of information aren’t powerful without something bigger behind them: a story.

Stories are persuasive. Stories stick with people longer, can have a bigger impact, and trigger an emotional response. It’s why we use a story like The Boy Who Cried Wolf to teach kids to be honest.

For data, stories give context, meaning, and help cut through the noise created by the Internet of Things.

But, there is danger in painting by numbers alone. By definition, stories are editorialized versions of the data. Data visualization can be inaccessible, oversimplify the complex, or tell the wrong story.

How can you avoid these pitfalls?

Lead with your story and let the data bring it to life. And the best way to do that is through custom data visualization—building your story from scratch, not from a list of options.

With a custom approach, you can position data accurately within the larger context, tailor it for your audience, and tell the exact story you need to tell. But only if you do it well.

Here’s a list of strategies to help you tell the story behind your data.

Make Your Data Visualization Sources Readily Accessible

Not to sound too much like a college professor, but…you will get deducted points for any source you don’t cite. Without any way for a reader to follow up with your data, your story can quickly feel illegitimate.

If your data is first-hand, providing an overview of your method for collecting will help a reader trust your story.

For example, a Bloomberg report about “What’s Really Warming the World?” cites its methodology for collecting the data it showcases.

First image shows visual graph, second image shows paragraphs of text

(1) A line graph showing the different levels of various factors on global warming, (2) Article finishes with detailed Methodology section

It includes the model they used, other projects they drew insights from, confidence ranges, and more. There are links for the user to learn more about each of the inputs that informed the numbers shown.

If it’s in reference to a larger report, providing a footnote source, with a direct link to that source, can let the “I-want-to-learn-more” user feel confident and comfortable.

As an example, let’s reference the Huffington Post article about millennial debt again. It’s designed to be very immersive and “hip.” But, to balance their design brutalist choice, they double-down at providing sources every step of the way.

Whenever they mention a statistic or create a graph, there’s microcopy directly paired with it to the source of the material.

Big statistics with text giving context and microcopy noting source

Microcopy used to attribute sources

Will a user actually investigate each and every statistic? Probably not. But the fact that it’s accessible for them to do it helps validate the credibility of the article.

Provide the Details Behind the Data, but First Summarize Simply

One of my favorite things to say about content is “less is more until more is more.”

For a casual reader—someone who is being introduced to your data, story, and thesis for the first time—you want to provide them with a high-level overview of the core concepts. Clear headlines and sections of data help users concentrate on one idea at a time, building their base knowledge for the deeper point you want to make.

For the informed reader—someone who is familiar with the concepts you’re sharing—you want to give them the ability to learn more. You want to quickly and easily bypass the overviews and dive into the details.

Data is complex, it’s important to contextualize and give users the information they need to understand the numbers.

One way to do this is with a simple headline pulling out the key piece of information from a data set.

For example, in an article about debt in the millennial generation, the Huffington Post provides detailed graphs as the user reads the article.

Summary sentence states, “Over 10 years, the typical ‘09 grad could earn up to $58,600 less than the typical ‘07 grad.”

“The Class of Oh No” graph comparing ‘07 and ‘09 graduates.

Each graph holds a lot of information, but rather than force the user to interpret it themselves, they provide a summary statement: “Over 10 years, the typical ‘09 grad could earn up to $58,600 less than the typical ‘07 grad.”

NPR takes a similar approach in an article that answers, “when are people working?” At the top of the article, they offer users a graph that can be filtered to compare jobs in different industries.

But, as the user continues to read the article, sections are pulled out and introduced with a short thesis statement that highlights how it affects the larger data set.

Graph shows food preparation and serving in comparison to all jobs, with a standard work day selected.

The paragraph above explains the graph below.

A curious user can explore themselves, and a user who wants to get to the “so what?” of the data can skim the detailed sections and cut to the quicker story.

Stagnating information and abiding by “less is more until more is more,” can be applied throughout the entire experience, and it can help you cater to different users and tell your story better.

Give Users the Power to Play

Digital data visualization has something that print can’t match: the ability to interact with the data. Don’t waste this opportunity.

Curiosity might have killed the cat, but it won’t kill your story: it’ll enhance it. Think of common questions your users might have about your data, and then brainstorm how you can let them interact to find those answers.

Ask: what is the main story to communicate? This should be the default view of the data with a clear summary (more on this concept later). But then, allow your user to engage and manipulate the data for their own purposes.

For example, an article from The Pudding, a digital publication (and a great source of data visualization inspirations, if you’re looking), investigates women’s pockets. Their story? “Women’s Pockets are Inferior.”

If you think about it, it’s something easy to get on board with. But, you may ask, “how much more inferior are they? Sure, women have smaller pockets, but compared to men, they also have smaller hands.”

The article answers this question—and rebuttals potential arguments against their thesis—with an interactive data set. Users can select different common pocket items—an iPhone X, a wallet, a women’s hand—and the digital experience responds, putting that item into a men’s and women’s pocket.

Shows comparison between iPhone X and Women’s hand, both fit at a higher percentage in men’s pockets.

Interactive chart to show difference between men and women pockets with a variety of common items.

Manipulation doesn’t need to be extensive: it can be simple and small.

That said, when you have the data, flaunt it, specifically when the story you’re sharing is complex and variable.

For example, the New York Times asked, “Is It Better to Rent or Buy?” That question has a billion different answers. It depends on your income and your city and how long you plan to stay and your mortgage rate options (and more).

So, rather than forcing users into one narrative, the NYT gave the users control to answer those questions—yes, all of them.

Summary on right includes top level number, then details costs after 15 years for the user

Graphs are all interactive and adjust the summary and recommendation

With data collected from across the country, they built a highly-interactive tool. Users can adjust 21 different filters, each impacting the answer to “should you rent or should you buy?”

For this story, it works. Someone who wants a specific answer is willing to interact with the filters to get it. But, be wary of building complex interactive elements into your digital experience: if you give users too much choice, you can create a paralysis.

You need to balance interactivity with stability to create the best story for your data.

Make the Visualization Cool, but Not Confusing

A designer can help bring your story to life—sometimes in a way that can imitate art. But it’s important to know that art is often up for interpretation, and, if you want to control the narrative your data is telling and communicate that message clearly, you should truly think about it as a design: not design as decoration, but design as strategy.

One way to achieve this is through metaphors. Simple, accessible visual connections can be a great way to tell a story while also servicing the data.

An example of this can be found in the work of Mona Chalabi. She often uses metaphorical elements to capture a specific idea, transforming it into a short story for her followers on instagram.

To showcase the problems with the U.S. Census, she created a graphic illustrating who gets double counted versus not counted on average.

Rather than use typical bar graphs, she used pencils, with the percentage not counted depicted with the eraser side.

It’s a simple, but powerful image, and layers a more serious implication into the post.

Pencils represent percent not counted and percent counted in a census.

Mona Chalabi pencil graphic commenting on census data.

Repetition is also a great tactic for simple, yet cool, data visualization. With repetition, users begin to predict the pattern. Then, small deviations become the thing that draws the eye: it creates focus for the user. You earn the ability to surprise.

The New York Times’ (NYT), interactive article about coronavirus deaths serves as an excellent example of repetition as data visualization.

It was published when the deaths in the United States had just reached 100,000. That number is big, and it’s hard to comprehend, let alone to humanize. So, NYT made a small icon to represent each person who lost their life to Covid.

Two people are darker in color with a caption describing a quirk about them, their name, age, and city.

NYT graphic article, Deaths at 10,828.

As a person scrolls down the page, the icons become overwhelming, but every page has a few icons that are a little darker than the others with a short fact about the person, their name, age, and city. The slight deviation in iconography draws attention, reminding the reader that there’s a person behind each icon, not just the ones shown.

With simple design choices, your data can be brought to life in a way that, yes is “cool,” but is also engaging, intentional, and well-paired with the story you’re telling.

The Data is The Supporting Actor. The Story is The Star.

Numbers help quantify the complex, often shadowy, elements in the world. With smart devices and artificial intelligence being introduced into almost every industry, the data we collect is only growing.

But numbers are only numbers without a narrative behind them.

To cut through the noise, it’s important to understand what story you’re telling, and then to tell it in a customized way that helps make the data more accessible.

For digital experiences, keep in mind:

  • Digital allows for interaction, which can drive engagement. Ask, “what questions will a user have about my story?” and see if you can answer them in either a small or big interactive way.
  • Cool is great, but clear is better. Use simple metaphors and design patterns to help bring your story to life without distracting from your thesis.
  • Less is more until more is more. You’ll have both newbies and specialists coming to your digital experience. Start off with an overview of your story, but also provide in-depth details for people to learn more and more about it.
  • Cite your sources. As a minimum, you need to include footnotes for your readers to be able to double-check your work. For more argumentative stories, more explanation will help bolster your credibility.

With these strategies at the forefront, you can transform your numbers into a compelling story: one that can build intrigue, inspire action, and spark change.