Telling Stories with Visualizations: Lessons from Data Journalists

by   |   September 25, 2012 1:09 pm   |   0 Comments

Image from NYT visualization on 100-meter sprint

A screenshot from “One Race, Every Medalist Ever,” a New York Times animated visualization showing how record-holder Usain Bolt’s performance compares to past Olympic medalists.

In recent years, journalists have been using data to tell stories in new ways.  From interactive data-driven visualizations that look at water supply shortages, building fire code violations or even wildfire danger, consumers of this information can often with a click of a mouse or swipe of a smartphone finger find out how a news story affects them.

Related Coverage

Learn more about implementing data visualizations.

       Read more»

Using data to tell a story that resonates is just as important in business.  “Data is hot right now, and people are expecting data-heavy presentations,” says Stephen Baker, the author of The Numerati, a 2008 book that examines the prevalence of data and how it’s used to study consumer habits. “If it doesn’t have data in it, people think it’s lightweight and it’s theoretical and they don’t take it as seriously. To be convincing, you need more data now than you used to.”

It turns out that journalists have a lot of expertise to share on the subject. Several top-flight programmers and data journalists lent their expertise for this story, and their advice was remarkably consistent: Know your data. Know your story. Know your audience. And keep it simple.

Takeaways from Data Reporters’ Notebooks

  • Data is the foundation. An original data-driven visualization with all the bells and whistles cannot stand on its own if the data is not compelling, or worse, inaccurate.
  • Refine your story. Mine the data to focus on one story. Using data that is not necessary, or not having a firm understanding of what the data reveals, can result in a muddled presentation.
  • Know your audience. Answer questions that matter to them. Why is this data relevant? What do they want and need to know?
  • Experiment. With data visualizations there are many different ways to present data. Consider alternatives and decide which one best tells the story.

Data Drives the Story
Lee Sherman, the co-founder and chief content officer of, a company that partners with publications and other companies to create data visualizations, says everything starts with the data, and the idea behind it.

“The data is way more important than the visual aspect. I don’t just mean hard numbers, but information and telling the story is important,” he says. “One of my biggest recommendations would be to make sure you nail the story first before you start drawing anything. Spend a lot of time up front finding your story. Come up with a narrative flow that reads like a story.”

Ryan Murphy, data reporter at The Texas Tribune, said he adhered to this approach in developing an interactive state map that took an in-depth look at the impact of drought conditions on the state’s water supplies.

The purpose of creating visualizations, Murphy says, is “making (data) easier to consume,” and having a firm grasp on the data and the story you’re trying to tell is a key component of that. “The data is where I start. That’s where I find my lead and what’s interesting and go from there.”

Of course, a major difference between producing data visualizations in a newsroom setting as opposed to a corporate setting is that a news presentation should let the facts speak for themselves. A corporation, however, has leeway to tilt the presentation to align with business interests. This is often reflected in how the data is mined – what’s used vs. what’s discarded – but also in how the data is ultimately presented.

Kevin Quealy of The New York Times graphics department illustrates this dichotomy when discussing the Times’ popular “Racing Against History” graphic, a visual that displays the results of every Olympic gold medalist throughout history in the long jump, the 100-meter sprint and the 100-meter freestyle. In the 100-meter sprint visualization, for example, Usain Bolt’s winning time is plotted on a track to show his time in relation to previous winners.

“Our principle in terms of what’s important is to really let the data and the content speak for itself,” says Quealy. “Whereas if you’re working for a business that say sponsored (100-meter silver medalist) Yohan Blake, maybe my entire focus about that graphic would be about Blake. We’d change the focus of the piece depending on what our business interests were. Those are very different principles at work. We’re just trying to see what we think is the most important (focus) based on the data that we have and of course what the news value is. So those are definitely different goals.”

While goals may differ, the same tenets apply in thinking about how to present data.

“In both instances you’d have the basic principles apply of knowing your data and understanding it,” Quealy says.

Focus On Your Audience
In news and in business, it’s important to stay true to the data you’re working with. Using that data, it’s essential to think through the story it tells.

Baker advises knowing the answer to some simple questions before embarking on a data-driven presentation: “What do we know from the data? What are we trying to find out? And are we using the data that’s really sexy to make a presentation, or perhaps the less colorful data that tells the truth that we don’t really want to confront?”

The reason, he says, is that “too often people are using data as a substitution for rigorous thinking.” Begin with understanding data and then move on to explaining what the story tells. The data, Baker adds, “might not be the sexiest thing around, it might just answer a simple question, but if it answers a simple question well it’s better than one that tries to answer a whole lot of difficult questions that muddles through.”

NPR fire forecast map for San Diego

An image from NPR’s fire forecast map for San Diego from Sept. 24.

Brian Boyer, who heads the news app team at NPR, which last month published an interactive map detailing wildfire dangers across the U.S., echoes some of the same thoughts.

“The thing I’ve been preaching lately is making your data useful,” he says. “The essence is that you have to know your audience and build a thing that specifically tells a story. Put yourself in their position. What does the audience want? Find that out and do that one thing well, as opposed to doing everything mediocre. Tell that one story and try to make it useful for people.”

Boyer cited the NPR fire forecast app as an example of making a data-driven visualization with a target audience in mind.

“You can make a map of all of the fires in the last five years and end up with a bunch of dots. You can even make the map playable, so you can start 10 years ago and hit play and watch dots pop up on the map. That’s pretty, and people are drawn to that,” he says. “But it doesn’t really give you what you want to know. When I hear America’s a tinderbox, I want to know if I’m safe today and if my backyard is going to go up in flames.”

Explore, Experiment and Revise
To make a presentation that fits well with the data, Quealy says, means you need to be ready to try out different ideas.

“The idea of knowing what your data is and knowing what works is because you’ve tried a few different things. The ideas [for how to present the data] come after you analyze the data, not before,” he says.

Quealy cites the Chicago public school strike as a recent example. The Times ended up making a sortable chart comparing salary levels and work schedules for Chicago and 115 of the largest school districts across the country. Chicago is the baseline, with other district figures represented in a bar chart compared to it.

“We didn’t just ask, ‘what are we going to make?’ We had to look at the data first,” Quealy says. If you were going to look at say how long the (Chicago) school day is compared to other school districts, you’d have to look at the other districts first and compile and analyze the data first before you’d know if a bar chart would work vs. other means. We choose the form to fit the data, not the other way around.”

Sherman, of, also mentions taking time to try different visualization methods as a vital step. He says that his designers always create non-visual representations of the data before beginning the design work, coming up with various sketches to determine what will most accurately and effectively tell the right story.

“Depending on what your data is and what story you’re trying to tell, it could lend itself to one format or another,” he says. The work trying out different approaches also can spawn ideas for future projects. “When you have a lot of data, you might want to start thinking about an interactive piece, and it’s also helpful to think in terms of re-usability” for datasets that come out in a similar format regularly, such as monthly, he adds.

Journalists also urge each other to follow one simple rule: focus narrowly on the story you’re telling. Data might seem like it can tell any story, but giving your audience too much to digest is counterproductive.

“I’d suggest people don’t make it hard on themselves,” Sherman said. “Don’t struggle to find ‘The Greatest Story Ever Told’ or something, because often times the story can be simple and it’s still going to be better than the PowerPoint.”

Ken Murphy is a senior writer for SAP Insider and Insider Profiles. He can be reached at

Tags: ,

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>