Data visualization in the age of “fake news”

Erase the labels from a bar graph or pie chart, and you’ll have no idea what it’s about. Influential data journalist Mona Chalabi sees that as a pervasive problem with our most commonly used data visualizations.

“It’s important that the visualization itself reflects the subject matter and not just the numbers,” Chalabi says. “A lot of visualizations strive toward this neutrality, which is quite bizarre. They’re trying to seem more objective, but they’re not using the visual language well.”

“If you take away the labels, a chart could be about anything. You wouldn’t know whether it was a chart about salmonella poisoning or the percent of people who say they’re happy on any given day,” she continues. “I want my charts, when you look at them, to immediately convey what the subject matter is and then to go into the numbers.”

Chalabi addresses issues of honesty, accuracy and accessibility in data visualizations in a free, public talk on Monday, March 27, from 7-8 p.m. in the James B. Hunt Jr. Library Auditorium. Her talk, “How Numbers Can Mislead (and how to avoid being misled),” will cover the imprecision of numbers and introduce ways to make honest and inclusive data visualizations that could help rebuild the public’s trust in numbers and facts.

Her appearance is part of the NCSU Libraries Presents “Coffee & Viz” series and is generously sponsored by the Eastman University Engagement Fund. While admission is free, tickets must be reserved in advance on Eventbrite.

Chalabi is a pioneer in the field of data journalism, having worked in many contexts, measuring data about things like humanitarian need, banking risk and corruption. She’s currently data editor at The Guardian US and has previously worked at FiveThirtyEight, the Bank of England, the Economist Intelligence Unit, Transparency International, and the International Organisation for Migration.

“My job is about using numbers to help inform people about the things they want to be informed about,” she says. “Like any journalist, I’m not really worth my job if I’m not honest. It’s the most important trait there is. And I think that being honest about numbers is particularly important and challenging right now.”

Although the term “data journalism” is new, Chalabi says that the field is old. “Data has always been a part of journalism. In the first issue of The Guardian, which was published in 1821, there was a table in the back about school access,” she says. “It was always understood that collecting numbers was a part of journalism, just like words are a part of journalism.”

“I think that the term ‘data journalism’ is being used now partly because we have so many tools to do data journalism, which offer a lot of new possibilities. But they can also offer journalists a negative temptation to overstate the accuracy of their numbers and the certainty of their predictions.”

Chalabi brings a pragmatic, holistic take to hard data, sometimes producing hand-drawn figures and animations that make her data-gathering process more transparent. “I’m concerned about how people overthink accuracy and certainty, and I think that the pure aesthetic of making something hand-drawn inherently has this element of uncertainty about it,” she says. “When you look at my charts, you don’t walk away with this statistic like 37.45% of people eat pizza for lunch—you see that about a third have said they eat pizza for lunch. All you really need to know is the ‘about a third’ element of it, because the methodology for collecting those statistics is actually so imprecise that the decimal place means nothing.”

“Another reason is that hopefully people can see my process a bit better, because there’s something satisfying about showing people one axis at a time, and then plotting the data, and then adding the labels, as opposed to giving them all the visual information at once.”