The deal with data

In high school, Chad Skelton hated math because he saw it as “detached from the real world.” Talking to his guidance counsellor, he learned that all university programs required grade 12 math … all but one: journalism. So Skelton became a journalist.

Things changed when he joined the Vancouver Sun in 1998. He worked with Freedom of Information requests and became “the math guy.”

In 2008, the Sun obtained a Freedom of Information request showing Vancouver daycares had risk ratings according to health inspections. Skelton wanted to do the story, but felt something was lacking. A few days later, the Sun held training on “Computer-Assisted Reporting” – now known as data journalism. He didn’t know much about data journalism, but after the training he felt a “lightbulb went on.”

Before, he had a general idea; now he could provide specific facts and a search tool.

Nobody in the Sun newsroom, he says, was diving into data journalism. “It was this huge world I could explore on my own. There were so many opportunities, it seemed like a no-brainer.”

Data is needed, he says, because it makes news hyperlocal. Take, for example, a crime map. A reader can look at what’s going on near their house, on their street, in their town. For Skelton, compared to other forms of journalism, this makes data journalism stand out.

Dive into data

So, what is data journalism?

Well, this is debated even in the data community.

For Jack Julian of CBC News in Halifax, trying to define data journalism is like trying to define medicine. Medicine is broad, with many subsections: surgeon, nurse, personal care assistant. Similarly, data journalism has subsections: data visualization, cleaning (identifying corruptions in data), scraping (extracting large data from websites into a file), mapping.

Data journalism also has many names: data science, precision journalism (dating to 1968), and computer-assisted reporting (dating to the 1980s).

Simply put, data journalism is using and analyzing numbers in a story.

Let’s say a journalist collects information on speeding tickets in a city for the last five years. Analyzing this data, it’s discovered that more speeding tickets are given out during holidays. This is data journalism.

Data is everywhere: the price of houses, sports statistics, where the taxes you pay are being spent. Data journalists make sense of those numbers, and frame them to be meaningful.

Chad Skelton now teaches data journalism classes at Kwantlen Polytechnic University in Surrey, B.C. He says students who suffer with “math-phobia” find data journalism not as intimidating as they thought. In fact, it’s fun.

“It all comes from the same desire that journalists have to explain and understand things,” Skelton says. “Data gives them another tool.”

He encourages all journalists to become comfortable with data journalism. Knowing how to crunch data, he says, means that “when someone comes across numbers on their beat, they don’t have to throw up their hands and assume that someone is telling the truth.”

He hopes more journalists will adopt data as a regular tool. But journalists, and journalism programs, he says, first need to recognize and prioritize numbers as important.

When Skelton began teaching at Kwantlen, data classes were an elective. Now, a data class is required. He says that when it wasn’t required, most students wouldn’t take it because the idea of data was scary.

Lack of confidence

John Wihbey has taught journalism at Northeastern University in Boston for five years, integrating data into his classes. “Journalists are famously averse to numbers,” Wihbey says, “but it’s also our bread and butter in certain ways.”

Some stories journalists cover – politics, health, economics – have a slew of numbers. By looking through data, journalists can find stories. So, journalists should love numbers, right?

In 2015, Wihbey conducted an online survey that proved otherwise. The survey was through the Shorenstein Center, a research centre at Harvard University. In the survey, 1,118 journalists and 403 educators participated, considering their skills and knowledge. The survey was open internationally, but most were journalists in the U.S.

Eighty per cent said it was “very important” to be able to interpret statistics, but only 25 per cent said they could do this “very well.”

The problem is clear. How can so many believe dealing with stats is vital to journalism, but not be able to do so?

Wihbey says the results were “unfortunate, but realistic.” Although he was a journalist before moving to the classroom, he didn’t “embrace” math until he was a teacher.

Journalism educators often fail to train students to deal with numbers, he says, and all journalism schools should require a data class. But adding data classes won’t solve the industry-wide problem. It’s also up to journalists to teach themselves, he says, through online courses and support.

“It’s okay to start from square one,” he says. “We’re all a work in progress, even the best data journalists. Whatever level you’re at, there are more skills to be gained.”

In 2017 a Global Data Journalism survey, opened to journalists worldwide, had 206 participants from 43 countries. Its findings were presented at the first European Data and Computational Journalism Conference by Bahareh Heravi, an Assistant Professor in the School of Information and Communication Studies at University College Dublin.

Ninety-one per cent agreed or strongly agreed that data journalism improves the quality of journalistic work. Eighty-six per cent considered themselves data journalists, but just 18 per cent saw themselves as experts.

Becoming expert often involves frustration, even among data lovers. “If I sometimes didn’t get flustered when working with data,” Alexander Quon says, “I wouldn’t be looking in the right place.”

Quon is an online producer and reporter for Global News in Halifax. He chases daily news in Nova Scotia and New Brunswick, while making time on the side for longer data projects.

He’s a recent graduate. His bachelor’s degree didn’t teach him any data journalism, so he completed a master’s degree in 2017 to specialize in investigative and data. He describes the experience as “going from zero to a hundred.”

He says the data tools he learned are what got him a job. His first big data story as a professional journalist was to investigate seven years of provincial inspections of Nova Scotia restaurants. The story took him a year to do, and was published as a Global News series.

He gets most of his data-driven stories through Freedom of Information requests or public data libraries. “Stories are hidden everywhere. You just have to be willing to go find the data.”

Compared to the U.S., Quon says, data journalism in Canada is less established and still figuring itself out. He believes journalists need to throw out the mindset that math doesn’t matter, and calls data journalism a “growing field.”

Data schools

A 2016 study, Teaching Data and Computational Journalism, examined the state of 113 university journalism programs in the U.S. The study called for universities to improve and increase their data-driven classes, to better the skills of aspiring journalists.

The 113 programs are reviewed by the Accrediting Council on Education in Journalism and Mass Communication, which includes applying numerical concepts as a core element of journalism education. Fifty-four of the 113 universities offered no data courses at all, despite this being an ACEJMC core value. And only eight universities offered a course in advanced data.

Charles Berret, who was then finishing his PhD at Columbia University and now teaches at the University of British Columbia, was a co-author of the study. He says that while data journalists are doing “the most innovative and impactful journalistic work year in and year out,” there is an overall lack of journalists equipped with the skills “or even inclination” to add data to their stories.

A million rows

On a Friday afternoon, Jack Julian sits at his desk in CBC Halifax. He scrolls through the spreadsheets of his most recent data story. It’s an investigation into the 178 ATV accident deaths in Atlantic Canada from 2012 to 2018. The data includes genders and ages, whether the deceased were intoxicated or wearing a helmet. He remembers a time when Excel only allowed 56,000 rows; now, it allows about a million.

Julian joined CBC in 1997 as an arts reporter. He wasn’t passionate about the arts, but like many graduates was desperate for a job. Today he does news reporting, sprinkling in data stories when he can.

Data journalism has changed a lot. When he first started, whenever he pitched he had to explain what data journalism was to his bosses. Doing spreadsheets was “ground-breaking.”

Will journalists continue hiding under the covers from the scary math monster in the closet? Alexander Quon of Global News in Halifax thinks data journalism is going to get bigger as governments become more secretive. “The only way you’re going to be able to pry stuff open,” he says, “will be looking through data. Numbers don’t lie.”

Julian is concerned about how data journalism will be funded. It’s expensive and time-consuming – things shrinking newsrooms don’t look at kindly.

“We have big companies like Google and Facebook accumulating data at an extraordinarily high level,” he says. “It’s important to have data journalism skills in order to hold those companies even remotely accountable.”

If journalists learn the tools, Julian says, data journalism will evolve in ways we cannot imagine. He says virtual reality could also be used to represent and interact with data.

One thing is certain, he says: “We are never going back.”

About the author

Madeline Biso

Madeline Biso is a student journalist at University of King's College. Her main interests are investigative and data-driven stories. When not...