Canada’s data journalists are using numbers to tell previously impossible stories. They want you to do the same
In 1994, Fred Vallance-Jones had been a journalist for 10 years, travelling across western Manitoba reporting stories for CBC Radio. Intrigued by a new and rapidly growing type of journalism called computer-assisted reporting, he flew to Ottawa to attend a training session organized by the Canadian Association of Journalists (CAJ). “They pulled out this mysterious database software called FoxPro, and I was amazed by what it could do,” he recalls.
Back home, he dropped $150 or so for a copy at his local Future Shop. “I then promptly realized that in two days I hadn’t really learned very much about how to use it,” he recalls. So the next spring, he drove to Columbia, Missouri, to attend a five-day boot camp run by the National Institute for Computer-Assisted Reporting (NICAR), a program of the Investigative Reporters and Editors (IRE) organization. There, he learned how to properly use FoxPro to “do the thing I actually wanted to do.” He was “blown away.”
When Vallance-Jones got home, he began putting his new knowledge to work. There had just been a deep recession, and the federal government was distributing billions of dollars in an infrastructure program. “It was basically throwing lots of government money at building bridges, highways, community centres, and the like,” he recalls. Curious about how the money was being spent in Manitoba—expenditures that were being decided by the provincial Conservative government—he asked the province for an electronic list of the projects. “They balked, but they gave it to me on paper,” he says.
Back in the office, Vallance-Jones entered the data meticulously into FoxPro by hand, cross-referencing it with voting results. “I was interested in, well, was there any political sort of favouritism in how the money was handed out? And lo and behold, there was quite a bias toward the governing Conservatives.”
“That was my first sort of data story.”
Vallance-Jones continued specializing in computer-assisted reporting, eventually leaving CBC Manitoba for The Hamilton Spectator before becoming an instructor at Halifax’s University of King’s College in 2007, where he currently directs its journalism program department. Today, he could be considered a trailblazer of Canadian data journalism. “I always say that data journalism made my career. If I hadn’t gone and done that boot camp, I’d probably still be somewhere doing the afternoon news and not have had the career I had.”
Working with data, Vallance-Jones says, is “the ability to look at things from a 30,000-foot view—the big picture—at the same time as looking at things one by one. It’s made things possible we couldn’t do before.”
Matt Frehner, managing editor of The Globe and Mail’s products and platforms department, which includes the paper’s data team, elaborates. “The storytelling potential of data, coupled with the technology that allows you to interrogate big, big datasets in a much quicker way than you would ever be able to as a human being just opens the door to completely different types of journalism.”
That journalism has caused shockwaves in the industry. Many cite the Globe’s 2017 “Unfounded” series as a watershed moment for Canadian data journalism. The #MeToo-era investigation revealed that Canadian police dismiss one in five claims of sexual assault as unsubstantiated. The series sparked an uproar. The granularity of the data gathered by reporter Robyn Doolittle and the team meant readers could look up the numbers in their own home cities and provinces. Conversations were had, policy changed. “Unfounded” was lauded, both at home and internationally, and the Globe subsequently ran dozens of related stories.
Since “Unfounded,” data-driven investigations, from local to country-wide, have continued to have an impact on Canadian journalism. In 2021, investigative reporters Tom Cardoso and Tavia Grant, also with the Globe, revealed that the Catholic Church in Canada—despite claiming its coffers were too modest to pay the financial compensation it had promised to Indigenous survivors of its notorious residential schools—had over $4 billion in net assets, sparking renewed calls for reparations. The next year, a four-part investigation by Robert Cribb at the Toronto Star and the Investigative Journalism Bureau found that some Ontario drivers with medical conditions were unjustifiably having their licences suspended, wreaking havoc on their lives. In 2023, Francesca Fionda, a journalist and now also director of newsroom development at The Narwhal, revealed in The Tyee that B.C. residents who flee environmental disasters are displaced for approximately more than three weeks—an amount of time the province’s support responses were not designed to deal with.
Philip Meyer of the Detroit Free Press is credited by many as the first journalist to employ computing to assist with his reporting. In 1967, in the aftermath of the riots that had erupted in Detroit that summer as Black residents protested systemic racism and police brutality, Meyer analyzed data using a mainframe computer to investigate and dispel racist claims about the uprising that had been coursing throughout the media. By the 1980s, as desktop computers became more common, a few journalists—predominantly American and many of them members of the IRE—began to view data as a significant source and tool. David McKie is a contemporary of Vallance-Jones and is another early leader in Canadian data journalism. Now deputy editor of Canada’s National Observer and a data journalism instructor at Toronto Metropolitan University, Carleton University, and University of King’s College, the projects were “groundbreaking,” he says. “The IRE was made up of real data nerds, who were off in the corner, doing stuff.”
It took a little longer for a data nerd class to emerge in Canada, though by the late 1990s, the Globe, CBC/Radio-Canada and the Toronto Star were using computing for some reporting. In 1999, the CAJ created an award for computer-assisted reporting, which in 2012 was rebranded as data journalism. “But,” says McKie, “there was always a very small community in Canada compared to the United States.”
One early member of the community was Valérie Ouellet, now a reporter with CBC’s Investigative Unit. In 2014, she took a leave of absence to complete an accelerated, data-focused master’s degree at King’s College, where she was taught by Vallance-Jones and McKie. When she returned to her newsroom, “ they didn’t know what to do with me,” she says. “It was a tough sell in Canada in 2015, trying to pitch data journalism in newsrooms.” Speaking of his own experience in that time, McKie recalls, “Our executive producers couldn’t wrap their heads around data. And more to the point, they really had no idea how to double-check your work. I think that gave them a lot of insecurity. And because of that insecurity, they didn’t have the same kind of confidence to say, ‘Let’s go ahead with this.’”
Then, in 2017, in the same month that Ouellet was hired in her first data reporter position at CBC, “Unfounded” came out. She noticed a shift: where people had previously thought of her as a wizard dealing with nebulous, nerdy stuff, there was now an understanding of the purpose of data journalism. Enthused by the excitement around “Unfounded,” CBC fast-tracked and ramped up its own data investigations. She says, “Unfounded was huge for data journalism in Canada. For people like me, it meant that my boss suddenly was like, ‘How do we do the next Unfounded?’”
While the pool of Canadian data journalism practitioners remains low in comparison to that in the United States, McKie believes the ranks are growing. “There’s so many open data portals now that it’s hard to be a functioning journalist and not know how to download a simple dataset,” he says. At the 2024 CAJ conference, he says, the sessions on data journalism were full. “I mean, they’re small, but they were full.”
Today, CBC/Radio-Canada and the Globe remain among the chief Canadian news organizations practicing data journalism. At the Globe, what began in 2014 as an interactive news unit has evolved into a larger team of data and visual journalists, where three data editors work on a 10-person data and digital storytelling crew. “We are a small team that prides ourselves on punching above our weight, often with far fewer resources than the heavyweight publications on the international stage,” says Danielle Webb, deputy visuals editor for data and digital storytelling. “While we don’t do the work for awards, we do regularly find ourselves in the company of industry giants like The New York Times and The Washington Post during awards season.”
“Journalists do themselves a disservice saying, ‘the reason I got into journalism school is because I suck at numbers’”
David McKie
Chen Wang joined the Globe in 2018 as the first data editor hired for today’s dedicated data journalism team. Typically, she works on data components of projects in collaboration with reporters. Her bylines have appeared in pieces about housing and underused public land, abandoned gas wells, hate crimes, vaccine uptake, and gender inequities in the workforce. “We’re more like general assignment reporters; we walk on all kinds of beats,” she says. Recently, though, she’s started conceiving and writing her own pieces, too. “I want to have a say in the final story,” Wang says. “I want to have some more original ideas.”
“Some might call it too nerdy. It’s very nerdy,” says Wang’s colleague and fellow data editor Yang Sun. “But based on our analysis of the readers’ data, Canadian readers, or at least the readers who look at the Globe, they welcome really nerdy, analytical, in-depth analyses.”
Wang adds, “In the era of misinformation and disinformation, it’s very important to protect our credibility and we hope data can contribute to that.”
It was only a decade or so ago that CBC/Radio-Canada hired its first dedicated data journalist. A couple of years later, three more were added, including Ouellet. Today, around half a dozen full-time, data-specialized journalists work across the country in various divisions, including the Investigative and News Labs teams. Using data, they have produced stories on everything from Ontario police officers being paid significant amounts of taxpayer money while suspended to the unequal access to sports and recreation facilities that exists across Montreal. Its award-winning Climate Dashboard, an interactive web page comparing current and historical weather conditions, updates itself in real time using the latest data from the Meteorological Service of Canada.
Naël Shiab, a senior data producer with News Labs who helped develop the Climate Dashboard, says he kicks off his data journalism with a question. Where can you, specifically, afford to rent? How did domestic flights contribute to the spread of COVID-19 in Canada during the pandemic? Who lives in your city’s worst heat islands? The end product—for which Shiab makes sense of huge numbers of data points by way of coding—is delivered in an interactive online format. “We try to make it talk to you, in the story,” he says. He calls himself a computational journalist, explaining, “When you start working with a lot of data, really quickly you realize that you need this computational, computer science knowledge to actually deal with all of this crazy amount of information.”
Sun agrees. “Being able to code is a default skill set of ours, and we want to adopt machine learning and computer science and to use large language models or statistical models to do more advanced analysis. We are still at the front line of using technologies.”
An increasing number of journalists on the front line of Canadian data journalism are women like Wang, Sun, Fionda, and Ouellet. In Alfred Hermida and Mary Lynn Young’s 2017 academic study, “Finding the Data Unicorn,” 15 out of the 17 Canadian data journalists they interviewed were men. Sun, who previously worked in the States, recalls going to two NICAR conferences and remembers who was grabbing the most attention. “It’s not only male—it’s white-male dominated, kind of like the tech industry.” Now, though, she says, “in Canada, just based on my community, I can see a big representation of women.”
Ouellet says that in the 1990s and even into the early 2000s, there were so few women at NICAR conferences that “they could all go for drinks together at one table.” A few years ago, she was at a NICAR conference where the female attendees rented out a whole bar. According to McKie, NICAR’s yearly conferences now draw thousands, and attendees are no longer solely “data nerds” but regular reporters who have recognized the need to better understand data. “That speaks volumes,” he says. “Are we drawing that many people, even proportionately? No, no. Not even close.”
“Sometimes,” says Wang, “you feel a little bit isolated in the Canadian media landscape, because there’s only like one or two people in each media outlet, and it’s really hard to find connections.” Across the border, McKie notes, “You just have a larger slice of the journalistic ecosystem that is more conversant with numbers than we are, with data. I’m talking about journalists of all ages, especially young journalists. That’s what’s still holding us back—we don’t have that kind of data literacy that we need within the journalism ecosystem.”
Indeed, since data journalism arrived on the scene, some have been arguing that every journalist—whether specialized in data or not—should be able to open up a spreadsheet and do basic analysis. “We were saying that in 1995,” says Vallance-Jones. “Everybody who’s seriously taught data journalism has said that forever.”
The idea of widespread data literacy, however, runs up against the notion that journalists are not numbers people, a stubborn stereotype in the industry. In newsrooms across the country, reporters who are otherwise very competent journalists confess—sometimes playfully or proudly—to being weak when it comes to numbers, something that makes McKie “want to pull my eyeballs out.” He says, “journalists do themselves a disservice—they always have—by saying, ‘well, the reason I got into journalism school is because I suck at numbers.’”
Wang and Sun’s team has held workshops and training sessions on basic data tools for fellow Globe colleagues. Though journalists aren’t always immediately adaptive to the process, Wang is still grateful for the coworkers that show up and show interest. “Most tend to say, ‘I have a dataset, can you help me?’ They do not want to learn how to catch the fish,” she says. “People are happy to stay in their comfort zone.”
In some ways, the data journalism learning curve has become steeper. “There are many journalists who are self-taught. It is possible,” says McKie. “But now the AI component is becoming so sophisticated that it’s virtually impossible, if not impossible, for anyone to learn how to do advanced level coding on their own—unless you go to school for it.”
On the other hand, there is much you can do using a simple spreadsheet. “A lot of what I hear is, ‘I don’t know how to code, so I don’t know how to do data, so I can’t do a data story,’” says Ouellet. “That is not true. You can absolutely break one million stories, or a decent amount of stories, just knowing how to do things in Google Sheets.” Carly Penrose of the Investigative Journalism Foundation agrees: “The ceiling does not exist. But the floor is not that high.”
The data-savvy journalist likely also has a leg up in today’s job market. “It’s one of the only parts of the industry that is actually growing,” says Ouellet. “The industry is actually hiring people as data editors, as data journalists.” Cecil Rosner, the Investigative Journalism Foundation’s managing editor, who hired Ouellet while he was at CBC, says data literacy is a skill he looks for in potential candidates. “I’ve been involved in a lot of hiring over the years. I’ll interview someone, and they’ll often say, ‘I’m a storyteller, that’s my main job,’” he says. “And I think, how do you know whether the story you’re telling is true? Is that an anecdote that you’re telling? Maybe you find two people that have had an unfortunate experience in a hospital waiting room. Do you represent it as, ‘This represents the state of hospital emergency rooms in Canada right now,’ or do you try to find the data?”
However, data journalism takes time and human resources. Typically, you must get your hands on the data, which can be a months-long ordeal in Canada for any data that is not open source, and then you must clean, analyze, and turn it into accessible maps and charts. For deep-dive, top-notch data journalism, widespread collaboration—as with any big project—is essential. “Unfounded” took 20 months from beginning to end. Almost 250 freedom of information requests were filed, and data from nearly 900 police jurisdictions were painstakingly analyzed. Twenty-seven people across the Globe had a part to play at one point or another.
At smaller shops, staff are already stretched thin. “It makes it very tough when, as a reporter, you’re expected to wear many hats,” says Ben Waldman, an arts and life reporter with the Winnipeg Free Press. And Tyler Harper, editor of the daily Nelson Star in British Columbia, says, “We’re a small newsroom. We have to be jack-of-all-trades. I expect this is probably the case for almost every newsroom in Canada. Because we’re not The Globe and Mail or the Toronto Star.” Harper also wonders how relevant it is to increase data elements in the stories they do. “Ninety percent of stories I work on do not require me looking at spreadsheets or statistical reports. The 10 percent I do I love, but it’s not an everyday piece of community reporting.”
McKie, however, argues that data journalism doesn’t have to be a luxury for smaller newsrooms. “I would say that, if anything, data could make their job easier. That seems counterintuitive, but you just download a little dataset, and boom. And if it’s not there, due to the open data concepts, you can ask for that data.” Vallance-Jones makes a similar point: “If you don’t do this, then you’re missing opportunities to find stories. You’re stuck with the press release. You’re stuck waiting for the media relations officer to get back to you two weeks later. A database can give you the answer in two minutes, once you have the database.”
Still, data cannot be seen as a panacea. Vallance-Jones cautions that the mass analysis of data has also introduced the possibility of making massive mistakes. “Data is not always accurate, it’s not always clean. Sometimes there’s duplicated records, sometimes there are things that are just plain wrong.”
Fionda says: “You always have to be questioning the information that you get, no matter who it comes from, and make sure you understand the process of how it was collected.”
The possible threat of “massive mistakes” is why data-specialized journalists often rely on others to double-check their work. “I just want somebody to look over my shoulder,” says Sun. “With numbers, you always want to make sure it is correct, and somebody else can help validate your results.”
To be transparent with readers, some also publish the research process along with their stories. “When you’re criticized about your story, you can say, ‘Hey, everything is public. Tell me which part is wrong. I have nothing to hide,’” says Shiab. “This full transparency I hope can help with public trust.”
There’s another aspect of data journalism that some practitioners are concerned about: numbers getting in the way of the story. “We are not doing nearly enough narratives—we are seeing data used as an end in itself, rather than a jumping-off point for strong journalism,” Sarah Cohen, Knight Chair in Journalism at the Walter Cronkite School of Journalism, told the Global Investigative Journalism Network last year. “I’m a judge for data journalism awards, and I can tell you we throw out 90 percent of the submissions because they are great data exercises, but not great journalism. We get to do both—but if we forget to do the part that we’re good at as journalists, then what’s the point?”
Any dense, number-heavy story that’s forgotten characters, scenes, and narrative, forgotten that it’s journalism, is “not an indictment of data journalism. That’s an indictment of journalism, period,” says McKie. “It’s just bad storytelling. We have this kind of idea that just because a story is data-driven, it has to be about numbers and it’s going to be boring. No. It’s about people. The number is just the vehicle to get there.”
side bar:
Anatomy of a Data Journalism Story
Unveiling the real size of the Catholic Church in Canada’s coffers
In May 2021, the discovery of unmarked graves at the former site of the Kamloops Indian Residential School, located near the city of Kamloops, British Columbia, shook Canada and made national and international headlines. The Tk’emlúps te Secwépemc First Nation had surveyed the grounds and identified around 200 potential burial sites of children who had attended the school. What followed were renewed calls for the Catholic Church in Canada to pay reparations to survivors of residential schools, about 70 percent of which the church had owned and operated.
For Tom Cardoso and Tavia Grant, award-winning investigative reporters at The Globe and Mail, questions about the finances of the church were piling up. In an agreement negotiated in 2006, it had pledged to pay $79 million, which included making its best efforts to raise $25 million toward residential school survivors. But, by 2015, the church had raised less than $4 million, claiming that fundraising results had been weak, and in a controversial move, Prime Minister Stephen Harper’s federal government released it from its promise. “There had never been an answer as to why they didn’t pay,” says Cardoso. “They were claiming that they didn’t have the money. But why was that? Why were they released, and how much money did they actually have while they were making these claims?”
Within a week or two of the news breaking, Cardoso and Grant had decided they needed to investigate the net assets of the Catholic entities involved. “This is where my data brain, I guess, kicked in,” says Cardoso. “Catholic churches are all registered charities. And if they’re charities, their finances have to be public. So, we can just request every church’s financial statements or their summarized financial records and figure out for ourselves how much they’re worth.”
With the documents at their fingertips, Cardoso and Grant completed the first-ever analysis of the country-wide net assets of the church, including cash reserves, investments, property, and other holdings. In the end, Cardoso only did one preliminary calculation, landing at an estimate of a staggering $4 billion. Later, Charity Intelligence Canada, which specializes in exploring charities’ finances and their accountability and transparency, calculated the final figure. Cardoso had not been far off: in Canada, the church’s net assets were a minimum of $4.1 billion, or about 1,000 times the amount of money it had raised.
The 3,500-word, quick-turnaround investigation published on August 7, 2021, wove numbers together with a narrative, which included the voices of many people with a stake in the story. “Part of my reporting was talking to survivors, talking to the groups that were sitting at the table with the Catholic Church during the negotiations for the settlement,” Cardoso says. Though he didn’t end up doing the final calculations, Cardoso believes his advanced data literacy helped him understand how to get the story. “I know how this information is collected,” he says. “I know that it has to be public. And I know how to get it.”
In 2022, Cardoso and Grant shared a Michener Award with Jason Warick of CBC Saskatoon, who had broken the story that the Catholic Church had paid just a fraction of the money it had promised in 2006. “These two powerful investigations,” the awards committee wrote, “prompted a landslide of change, from a national apology from Canadian bishops and a renewed $30-million fundraising campaign for healing and reconciliation projects to a meeting with Pope Francis in Rome this spring and an historic apology to survivors.” The story was not over: Grant, Cardoso, and the Globe subsequently published several related articles, covering the developments.
Often, stories where data is an element have a larger effect on creating policy change, says Valérie Ouellet, an investigative reporter with CBC/Radio-Canada. “Once you have the number, it’s very hard for people to look away. It’s very hard for authorities to pretend it’s not happening.”
“We could’ve done that story purely anecdotal, about people who think that the church hasn’t done what it could to offer reparations,” says Cardoso. “But I don’t think that story would’ve had the same impact as a story that says all of that—but also says that they were saying they didn’t have any money when they had billions of dollars.” —Livia Dyring
About the author
Livia is in her final year of the Master of Journalism program and works on the Review’s senior editorial team. She is an associate producer at CBC Radio and has freelanced as a science writer. Livia is interested in art, science, current affairs, and stories that take a big perspective on the long, ongoing journey of humanity. When she’s not working, she loves listening to music, visiting art galleries, and spending time in nature.
