In November 2018, CBC Toronto released “The Implant Files,” a major investigative initiative about what happens when medical devices are defective, in conjunction with the Toronto Star and Radio-Canada. Valérie Ouellet, a senior data journalist with CBC, was on the project for about a year, with partners at Radio-Canada working on it since February 2018. Another partner David McKie, started exploring the concept over three years ago. The result, after months of communication with Health Canada: 45 columns and 350,000 rows of data, creating a table of 160,000 adverse incidents concerning medical devices and implants plus accompanying stories for web and television.
The biggest challenges with “The Implant Files” were figuring out how all the pieces of data fit together, as well as working out how to code the interactive database so that users could search to find names of devices, side effects and more.
For members of the CBC Toronto team, this was a small data set compared to what they usually work with. In November 2018, Ouellet and her partner, William Wolfe-Wylie, a CBC senior developer of news interactives, spoke at Toronto’s Data Driven 2.0 conference. The topic: Ticketmaster, an investigation in which they worked with over two million rows of data.
Wolfe-Wylie, who is a largely self-taught coder, spent the better part of a month creating the searchable engine.
“One of our big problems for a large–audience story like this, because [it’s] the CBC… there are hundreds of thousands, if not millions, of people who could have access to this and be performing searches,” says Wolfe-Wylie. “Writing efficient code and using more efficient techniques in writing code becomes really, really crucial when you’re writing programs that are going to be accessed by that many people.”
In the end, the project was comprised of a team of 24 journalists, with five journalists working on interactive elements and over 250 journalists from 36 countries consulting and sharing research.
“[Our team is] really hyper–specialized in [data], but there’s a number of other reporters across our network who do have data skills and use it in their daily reporting,” says Ouellet.
CBC is not the only news outlet to pour resources into a data team. The February 2017 “Unfounded” investigation from the Globe and Mail is chief among Canada’s data journalism triumphs. The culmination of 20 months of inquiries and reporting led to a parallax scroll with intricate and interactive maps, each created by the Globe’s visual team, alongside journalist Robyn Doolittle.
Matt Frehner, the Globe’s head of visual journalism, works with different departments to create visuals for print and online. His team is made up of people who code, analyze data and develop visuals, but the data is what they always return to.
“The data is really the whole story that you’re telling,” he says. “And structuring data in order to answer questions for the reporter.”
Not everyone on the team had a strong background in data or visuals, but learned throughout their time at the paper, says Frehner. “It’s a bit of a mix—like some people have come up to the newsroom and slowly evolved in that way and some people were hired with that specific skill set in mind.”
Without these people, some stories would take much longer to report, or be impossible altogether, says Frehner. “Just a human being would take years and years to do that kind of data analysis, but if you’re using software like Python [a coding language], structuring a data set that you can analyze makes things much quicker.”
CBC investigative data journalist Roberto Rocha spends his days coding and analyzing data that acts as the basis of stories, but he doesn’t think every journalist needs to know how to code: “For someone like me, who deals with data, coding is an incredibly powerful tool.”
Coders can, for instance, scrape websites for data, create bots that do sorting and create interactive visuals. For Ouellet and Wolfe-Wylie’s Ticketmaster project, they created bots that would go in daily and scrape ticket–selling websites to track seat locations and prices. Without the bot, they would have had to go in and manually follow the data themselves, which would have been time-consuming and labour intensive.
“These skills allow you to merge that investigative reporting with the more traditional boots-on-the-ground, getting-the-sources-lined-up kind of reporting,” says Wolfe-Wylie. “It adds an extra layer of verification and an extra layer of inquiry on top of journalism.”
“And I would say an extra layer of credibility,” says Ouellet. “Your questions are much more direct.”
Wolfe-Wylie equates data skills to superpowers, saying they make him and his team resources in the newsroom, allowing them to help fellow reporters.
In a field where a journalist must be adept at just about everything, it might be worth learning how to code and interpret data because it can do so much for reporting.
“It helps people make sense of very complex issues. Data has this has certain power. It has authority to it,” says Rocha.
Not only does data have authority, but those skills can make you a better at your job.
“I’m a lot more structured than I used to be in my research,” Ouellet says. “And a lot more systematic. I think it makes me overall a better journalist.”