Since the start of the pandemic, data journalists have had the difficult task of tracking massive amounts of inconsistent, ever-changing information around the coronavirus and wrangling it into visually appealing, digestible designs—on deadline.
In an email conversation with multimedia journalist Tiên Nguyễn, four data journalists covering COVID-19 discuss the most challenging aspects of the gig and their approach to producing graphics that are accurate, informative, and crucially during the uncontrolled pandemic, actionable. “A good [coronavirus] tracker exists to help people make decisions. At a local level, it helps people decide how to live their lives. And it helps policy makers decide when it’s time to close or reopen,” says data journalist Chris Canipe.
The reporters who participated in the discussion are:
Chris Canipe, data visual journalist at Reuters
Emily M. Eng, graphics editor at The Seattle Times
Jasmine Mithani, visual journalist for FiveThirtyEight
Aaron Williams, data reporter at The Washington Post*
* Editors’ note: Between the time he took part in this piece and the time we published it, Aaron Williams was hired as a senior visual engineer at Netflix. Congratulations, Aaron!
Tiên: What has been your biggest challenge in designing visualizations related to the COVID-19 pandemic and how have you managed this issue?
Chris: Uncertainty! The data for cases, deaths, and testing is messy and inconsistent. Some states include probable deaths in their official tallies and others do not. Some states have decided to include probable deaths after not including them before, which leads to a big spike in the data as all of those cases get added on the same day (*cough*, New Jersey). I built and maintain our state tracking page at Reuters, and I’m constantly afraid we’re going to misrepresent the numbers and imply big changes that don’t exist.
One way we manage those issues is through transparency. If you look at the “about the data” section at the bottom of any tracker, you’ll see a long list of caveats. At this point, the more caveats I see on a page, the more I trust the source. It says they’re aware of the inconsistencies and are doing their best to account for them.
We also have a responsibility to address those questions through reporting. I work with a reporter who writes a weekly analysis based on the data in our tracker, and each week, we identify big movements and anomalies and try to reach out to state health departments directly. Visually, rolling 7-day or 14-day averages help emphasize trends, which are really what matter. Rolling averages smooth out the noise and tell you what’s important: Are cases rising in my area, and how fast?
Aaron: Adding to that, I believe conveying uncertainty in the data—already a difficult concept in data visualization—made our jobs even harder. Because there was no singular national (and in many cases state or county level) resource for the COVID-19 information at the onset of the pandemic, the trackers created and maintained by The Washington Post, Reuters, CalMatters, and others very quickly became the go-to resources for this data.
This meant we all needed to be very careful about how we presented what we found. Many newsrooms accounted for this uncertainty by providing detailed and transparent methodologies about the inconsistencies in the data somewhere on the tracking pages, as Chris said. But to be frank, the reality is many people will view our visualizations without ever taking time to understand the limitations of what we’re showing.
This has always been a challenge with data visualization, but this specifically made the task of visualizing virus-related data all the more difficult because it’s important readers don’t incorrectly infer what’s happening in their county or state. We’ve all had to do continued reporting to make sure we didn’t misinterpret the data. I think Chris’s team’s approach to do a regular, weekly analysis is smart because there are so many variables impacting the data being reported. Explaining these nuances to readers is key.
Emily: For me, the biggest challenge around creating visuals for COVID-19 is the ever-changing landscape of the information and data. Here in Washington, we were considered the first epicenter of the infection in the United States, and back in late February and early March there wasn’t much information, especially scientific information, on novel coronavirus outside of China. We were learning about coronavirus at the same time the rest of the world was.
So, in our early visuals, I kept it simple—what did we know and what didn’t we know. I stuck to trusted sources, like the World Health Organization, the CDC, and our local public health department. I compared COVID-19 to the common flu—something our readers did know and could relate to. When I was unsure about any information, I asked an epidemiologist, researcher, or medical doctor to help explain it to us so we could explain it better to our readers.
As more information and more science has come out, we continually update our graphics and we err on the side of overexplaining the information. If there is missing information or the data changes, we add a note to our graphics. We also now have an explainer on how to read all the different types of graphs and charts we’ve been publishing, and in it we [also] emphasize the reading of trendlines over daily numbers and are able to add more context.
Jasmine: Whenever I design a visualization, I ask myself a few questions about how it can be used. For COVID-19 visualizations in particular, I have a specific checklist I run through before publishing:
- What harm could this chart and article do?
- If someone screenshots this chart and shares it on Twitter, does it contain enough information to be accurate without the additional context of the article?
- How could someone use this chart as justification to defy CDC guidance or local social distancing orders?
One way I minimize the potential for misinformation—always a concern but now even more heightened—is making sure the head and deck of my charts are clear and erring on the side of verbosity. I make sure all text is describing only what the chart shows and doesn’t rely on exposition—in an article it could be decoupled from—to make sense. I don’t hesitate to add a longer footnote to a chart explaining the data limitations, and I add annotations to key points in anticipation of questions a reader might have.
As Aaron brought up, communicating uncertainty has been one of our greatest challenges. Our whole team is invested in conveying uncertainty in a way that adds context but also doesn’t lead to readers outright dismissing the data we do have. As a larger editorial strategy, we published a feature about all the layers of uncertainty that factor in to COVID-19 models early into our sitewide coverage. This laid the groundwork for future work by addressing the limitations of data straight-on and primed our readers to think critically about the dozens of projections and headlines they saw in the news from us and others.
Tiên: When you’re designing graphics that synthesize an immense amount of coronavirus data, which is generally very distressing information, how do you try to make sure that the real, human impact comes through in the graphic?
Aaron: It’s always difficult to strike a balance between data and people, but I think that’s the role reporting and a story can have. It’s important to be smart about color choices and don’t choose anything too insensitive—e.g., using red in a graphic about violence. But on top of that, it’s also key to take the insights you’re showing in the graphic and find real people you can connect your findings to. At The Post, we did this by creating an interactive map as a resource for readers to look up health disparities in their communities, as well as writing a full feature story on how these disparities played out in the early days of the pandemic in communities across the nation.
Chris: I work with data, and what I’m thinking a lot about these days is the scale of this pandemic. Whoever said “a single death is a tragedy, a thousand deaths is a statistic” was right. The number of cases and deaths has reached numbers that seemed like the unlikeliest worst-case scenario back in March. But somehow, we’ve adapted to this new reality. Not just in the news, but in the culture, too. The United States has not had a moment of national mourning. There haven’t been any national ceremonies. No somber speeches, no televised reading of names.
I think visual journalism can humanize this story by connecting the data to individuals. I thought The New York Times‘ “An Incalculable Loss” piece, in print and online, was pretty moving. It forced readers to acknowledge the scale. And it used names and single-line obituaries to remind us that these were people. These were lives that were rich and complicated and ended in a hospital bed without friends or family around to say goodbye.
But as a consumer of news, I’m not sure I’m looking to data viz to humanize the pandemic. I read a lot of stories about individuals—doctors, teachers, parents, and children—and their experiences. This American Life and The Daily did excellent audio documentaries about health care workers, patients, and families. Narrative storytelling brings out the humanity for me. It’s how I connect the data to the suffering. But in my own narrow role, I have not done a great job of connecting the data to people.
Tiên: At this point, many journalists like yourselves have been reporting on the pandemic for months, which must be emotionally taxing. What ways have you found to take care of yourself during this time?
Chris: We’re working from home, and we all have our own unique life complications with childcare, fears about job security, family obligations, loneliness, and mental health. How do you focus when it feels like the world is burning? None of this is normal. A journalist I follow tweeted, “I’m okay, but I’m not *okay*.” And I think about that a lot. My job is secure, my family is healthy, and I work with people whose work I respect and admire. But this is trauma. We are all living with trauma, and we’re all going to come out of this a little fucked up.
I worked on a police brutality piece in June that hit me pretty hard. Building the piece required watching cellphone video of police punching, kicking, shoving, and firing tear gas at protesters following the death of George Floyd. For a couple of weeks, I was tense. I was jumpy. I was sad and angry. And I didn’t really know why until somebody pointed it out to me. Watching these videos all day is traumatic. And it’s especially so in the context of this moment, because whatever it is about the world that we’re all trying to compartmentalize, it finds its way into your brain in real and visceral ways.
So how do I cope with that? I take long walks in the morning. I remind myself to listen to music instead of NPR or podcasts. I take days off here and there to just go hang out with my wife and 7-year-old son. We go on bike rides and play board games and Mario Kart. We have socially distanced dinners on my parents’ back porch. I cope by remembering to do things that have absolutely nothing to do with work.
Aaron: During this unprecedented and difficult time I’ve found solace in cooking. There’s so much beauty and fun exploring new flavors and ways to reinvent dishes you love to eat. I’m also an avid music lover and have been recording mixes of some of my favorite music. DJ-ing is a great way to escape from the world a bit.
Jasmine: I’ve been lucky, in some respects, because our election coverage has kicked into overdrive and I’ve shifted to mostly coding our forecast instead of working the coronavirus beat. I’m not immersed in COVID-19 briefing and research every day, and don’t have regular contact with researchers or first-responders.
That being said, even if I weren’t doing pandemic-related reporting, the way our lives have changed has taken a toll. While stuck at home I’ve been traveling to other worlds through books and video games. I deleted Twitter off my phone a month or so ago, and that’s helped immensely by curbing a constant barrage of information and exposure to other peoples’ anxieties. I set up a morning routine to calm myself before the workday; I set small intentions, like trying to tap into my creativity more. I don’t focus on the end, or what next week or next year will look like. I focus on the present and appreciate the platform I am granted in order to provide useful, accurate, and meaningful reporting to the public in these difficult times.
Tiên Nguyễn is a freelance science journalist, video producer, and reformed PhD chemist. Her writing has been featured in Chemical & Engineering News, Nature, Knowable, and more. Follow her on Twitter @mustlovescience.