Storytelling with data: tips for success

By Julia Belden

This is the 2nd in a series of articles by Knight Center students who attended the recent annual conference of the National Association of Science Writers.

On a chilly November Friday, a gaggle of journalists recently huddled around tables in a hotel ballroom , attempting – with varying levels of success – to filter Major League Baseball salary data on a Google Sheet.

The goal? Equip science journalists attending the National Association of Science Writers (NASW) annual conference with new and powerful reporting skills.

Crunching numbers and wrangling spreadsheets don’t conjure the mental image that most people think of when they imagine journalists, but the aptly named sub-field of “data journalism” does just that, and more.

Adam Rhodes, the training director for Investigative Reporters & Editors, led the journalists – including a group from the Knight Center for Environmental Journalism – through a step-by-step process to clean, filter  and glean insights from datasets.

NASW data journalism photo: Investigative Reporters & Editors training director Adam Rhodes shows workshop attendees how to organize data in Google Sheets. Credit: Julia Belden

While modern data journalists typically do their work via computer, using data to discover and enhance news stories isn’t a new phenomenon, Rhodes said. For example, an 1848 story in The New York Tribune tracked milage expenditures of members of Congress, and journalist Ida B. Wells was the first to collect and analyze statistics about Black lynching victims in the U.S. in her 1895 pamphlet A Red Record.

Rhodes showed attendees where to find useful datasets online – government websites and reports are always sure bets – and how to prepare, or “clean,” the data for analysis.

Data cleaning is essential, Rhodes said, as errors and inconsistencies in the dataset can influence the analysis.

Rhodes told the group not to let a lack of specialized computer skills get in the way of data journalism – while some data journalists are skilled in programming languages like R and Python, “I know plenty who live and breathe Google Sheets, and they do really high-level analyses.”

Using the MLB salary dataset for practice, Rhodes led attendees through a step-by-step process for quickly sorting the data for insights, like which teams have the highest-paid players.

For visualizing data in stories, Rhodes recommended online tools like Datawrapper and Flourish. Datawrapper is easy to use and generates beautiful visualizations. (It also has a free tier!)

“It’s really hard to make something ugly in Datawrapper,” Rhodes said.

For inspiration – and a bit of journalism fun – Rhodes suggested checking out publications like The Pudding, which publishes creative data stories like how the type of animal on a wine label predicts the quality of the wine, and an examination of all-women songwriting teams.

As in any journalistic endeavor, context is critical in data journalism, Rhodes said. Connecting with experts, especially ones involved in the creation of the dataset you’re using, is essential for making sense of the data.

And, of course, it’s critical for journalists to talk to the people whom the data affects. Data is only a portion of the story, Rhodes said.

“You will always, always have human sources in your reporting.”

Julia Belden