When it comes to data-driven journalism, less is often more
By RHIANNA JACKSON-KELSO
Special to the RJRC
April 9, 2018
Data journalists are rethinking how they present data-driven stories as it becomes clear that readers won’t spend time clicking through interactive visualizations, says Roberto Rocha, a CBC data journalist and educator specializing in data-driven reporting.
Rocha, who has worked on stories ranging from a Montreal street history map to a ranking of psychedelic drugs, says data journalists must accept that readers have a limited amount of patience for sorting through data.
“Readers are lazy,” Rocha said during a Mar. 27 webinar hosted by Geothink, a Canadian geospatial open data research partnership funded by the Social Sciences and Humanities Research Council. “They don’t want to work to get information. Basically, they just want to scroll.”
Rocha said the early tendency of data journalists to throw every scrap of data at readers and hope that they would explore the material has proven ineffective. Most readers simply won’t spend hours, or even a few minutes, exploring interactive websites overloaded with datasets, said Rocha, who characterized the practice as “an abdication of the journalist’s role.”
More recently, he said, data journalists have been returning to the basics, which include filtering information, uncovering the main story, and telling that story in an engaging way: “There’s a major shift that’s happening in data journalism that started a few years ago where [the way data journalists perceive themselves] has shifted from simply organizing data … to more of an active storyteller role.”
This new, more user-friendly approach is evident, Rocha said, in the New York Times’ March 19 use of an animated graph to visualize rates at which white and black men who grew up in rich families remain affluent or fall into poverty. Another example, a timeline visualizing 311 calls New Yorkers made relating to damage caused by Hurricane Sandy from the fall of 2012 to now, uses scrolling to drive home the long-reaching effects of hurricane recovery.
“Interactivity is not a panacea for engagement,” says Rocha, noting that if readers must search for information, they are more likely to lose interest in the story.
Analytics have played an important role in reshaping the story-telling approach used for data journalism projects. Rocha said data gathered by the New York Times to measure reader engagement with its interactive visualizations – which are costly and time consuming to produce – revealed an engagement rate of just 15 per cent.
In response to these findings, the Times refined its approach to minimize the work required of readers and place more responsibility on the shoulders of the journalists. The new direction was summarized in a set of three “rules for visual storytelling” in a talk given by Archie Tse, deputy graphics editor for the Times:
- If you make the reader click or do anything other than scroll, something spectacular has to happen.
- If you make a tooltip or rollover, assume no one will ever see it. If content is important for readers to see, don’t hide it.
- When deciding whether to make something interactive, remember that getting it to work on all platforms is expensive.
Although efforts by data journalists to more effective in their story-telling approaches, panelist April Lindgren pointed out that many journalists still need to learn the basics of data-driven journalism.
“The capacity of local new organizations to mine … data for stories is really lagging behind the growing availability of data sources,” said Lindgren, the academic director of the Ryerson Journalism Research Centre,. Most local journalists lack the skills required to collect data and pull stories from it, she said, and recent graduates who do possess these skills often find themselves overwhelmed with the workload once they are installed in local newsrooms.
Lindgren said more collaboration between newsrooms might be one solution to this problem and pointed to a recent project by fellow presenter Zane Schwartz to illustrate her point. Schwartz, an investigative journalist for National Post and Macleans, spent his section of the webinar discussing the political donations database he compiled using more than six million records gathered from every province and territory. Though Schwartz noted only a handful of people had downloaded the full dataset, Lindgren says the production and sharing of such databases could be one way to address the problems faced by local news organizations that lack the staff, skills, and time to explore data-driven journalism.
“Local journalists could go [into these databases], take the data, and make it relevant to their local audiences and do that work for the people who live in their communities,” says Lindgren. “I’m hoping … that more local journalists will see what [Schwartz] has done and recognize what a gift it is in terms of a story in hand.”
She pointed to collaborative data journalism models launched in the United Kingdom as possible models for Canada. The BBC’s Shared Data Unit, part of its broader Local News Partnerships Program, pairs experienced BBC data journalists with reporters from the local news industry to educate them about working with and reporting on data. Similarly, The Bureau Local, launched in March 2017 by the not-for-profit Bureau of Investigative Journalism, aims to work with British local media to use datasets to write local stories.
Recreating these types of initiatives in Canada, says Lindgren, could be critical to improving the state of Canadian local journalism. And the data suggests these improvements are necessary—findings from the Local News Research Project, which Lindgren co-founded, show that 244 local news outlets of all types have closed in 181 Canadian communities since 2008.