Published in


illustration by Jason Forrest

Where Is the Rush to Visualize the Public Health Crisis of Racism?

Data visualization designers can be anti-racist by raising awareness of inequities and finding ways to #VizResponsibly

For the last three months visualizations of COVID-19 case data — with all of its uncertainty and flaws — occupied the 24-hour news cycle and dominated home pages of every major media outlet.

We saw a rush to visualize COVID-19 case data by a wide range of analysts and designers. We talked about visualizing data responsibly. But most of the news stories I’ve seen over the past week are photos of the protests, not visualizations of the systemic injustices our Black community in America has faced. Photos are powerful and important, but so is the data that illustrates the disparities in our country.

Home pages for the Washington Post and New York Times as of June 5, 2020 3:25 PM.

As a public health professional who watched endless dashboards of COVID-19 daily case data get spun up across the data viz community, I still question how much that ‘real-time’ data influences our choices to follow recommendations like social distancing or wearing a mask.

In contrast, I’ve seen how sharing data around inequities can raise awareness and shift perceptions, like the statistics in the Data 4 Black Lives Statement of Solidarity with Minnesota and analysis of fatal police shootings by race. W.E.B Du Bois’s charts remind us just how long these inequities have persisted. Each chart can chip away at the layers of unconscious (or more overt) bias built up over years of social conditioning.

Data exposing inequitable access to education, housing, and other resources is out there. Racism is being discussed as a public health crisis. Many sources have data of better quality than COVID-19 case counts. But somehow it doesn’t seem to captivate our community in the same way that incomplete and flawed COVID-19 case data did. Why?

What can we do as data people

We need to bring the same energy and (more importantly) the same care that we rallied for around the pandemic as we find ways to be anti-racist in our data visualizations.

First, and most importantly, let’s . They are actively sharing knowledge, ideas, and needs. Listen to them and share their work.

Then, find ways to be actively anti-racist through your data visualizations. Being anti-racist can be more than just disaggregating data, though that may be one piece of the process. It requires that we actively think about our motivations, bias in datasets, and what stories are being told by whom.

What follows is a list of questions for data visualization designers thinking about their role in the current movement. Some are adapted from the considerations shared around COVID-19 (which still apply here) and others were shared by friends, colleagues, or based in past experience making my own mistakes — I’m still learning too.

1. Who can you amplify?

You can start by following this list and recommending additional people and organizations to be added.

Compensate them for their labor. If you don’t have the resources to pay them for their time, don’t count them out of the project: Make the ask, and leave it to their discretion to say yes or no. Be mindful of what you can learn on your own to be more informed on issues of race in America. Spend time reading and learning so you can be a thoughtful collaborator.

@askdrfitz is an excellent person to look to if you’re unpacking how these two issues intersect. Throughout the COVID-19 pandemic, she has been calling attention to ways that systemic racism has paved the path to the disease’s disproportionate impact on Black people in America.

2. What are your goals?

Is your aim information sharing or enabling action? Have you considered how the visualization or narrative could cause harm, particularly to those represented in the data?

Even if your intents are noble, if you receive feedback from a Black reader about your visualization, listen to them. I appreciated how many people sought input and advice from public health professionals, epidemiologists, and biostatisticians on visualizations of COVID-19 data. Now, the experts we need to listen to are those who have lived experience being Black in America.

Consider how you can use your time and visualization skills to explore and expose issues of systemic racism. Take a card from a site like Mapping Police Violence and dig deep into one specific issue in our country.

Center individual stories and humanize the lives lost. Look to artists and designers who have created visual stories since long before the most recent set of protests, like Mona Chalabi.

Edit: I originally included a different artist’s tribute to Breonna Taylor rather than centering the large body of work Mona Chalabi has done on these issues— functionally not following my own advice and sharing the work of a white woman rather than a woman of color.

Instead of exploring systemic issues or individual stories, you could create tools to help people find Black-owned businesses or social justice organizations. Looking beyond donating to causes, changing our buying behaviors to support Black-owned businesses with our dollars is far more impactful than a black square on Instagram.

where to learn more, donate, amplify, or otherwise support anti-racist movements. Give your readers something they can do to follow up on what they learn from your visualization.

3. What data to use?

with the same depth that a journalist would take when vetting a source. Who and what isn’t represented in the data? Here are seven big considerations to start with: the We All Count Data Equity Framework.

. With the publication of books like Invisible Women and Data Feminism, we saw an outcry in the ways big data (and sometimes small data) does not represent all genders. When we don’t have disaggregated data available, you cannot always assess how representative the dataset is of the broad population or subgroups.

If you’re pursuing a data visualization project and cannot find race disaggregated data on your topic, that absence may be a story worth understanding more deeply. The data may be masked to protect the identity of those represented, particularly in small samples, or not be collected with that level of detail. We may not be able to analyze and understand how groups are disproportionately impacted by different issues without those breakdowns or more detail on the generalizability of a dataset.

In March, fewer than half the US states were reporting COVID-19 data disaggregated by race. Today, in early June, only two states continue not to report this breakdown. Public pressure seemed to be a key factor in the change in behavior on the part of the states who shifted to release disaggregated data.

4. What design choices are you making?

with stock photos and ensure you have someone’s consent to be featured in your visualization. @pexels actually includes conditions on this use in their terms and conditions (see Section 6).

If you’re sharing visualizations that include photos of protesters, think twice if they identify someone and you don’t have their consent to use and share the image. See some of the photography “don’ts” here.

Many sites primarily feature white photo subjects. Seek out more diverse photos, like on sites like or TONL.

Many visualizations of COVID-19 were not accessible to those with visual impairments or cognitive disabilities. Design with accessibility in mind. Here are five tips to start with from @AmyCesal on the Storytelling with Data blog.

Words can have coded meaning. Be open to feedback, particularly if it comes from someone represented in the dataset. Seemingly simple things like capitalization matter: here’s why using Black with a capital B is important, for example.

, particularly in borrowing from Black culture. For decades, Black culture seems to have been something to consume but not engage with. It is not ours to appropriate as white designers.

5. What story does your visualization reinforce?

For a lot of topics that might be addressed in these kinds of visualizations, your readers will need more context to make sense of the visualization. Given how underrepresented Black history and discussions of race (and colonialism) can be in our education system, we should not assume our readers have that knowledge. Find recommended reading, add the links to your visualization, and prioritize sharing sources from the people or communities represented in your data.

How could they be misunderstood? Are you (perhaps unintentionally) reinforcing a stereotype by not considering the context within which data was collected or other peripheral sources you need for your analysis?

With COVID-19, we saw headlines pointing to the disproportionate burden of disease in Black communities. It’s not enough to state that there were more cases or deaths: We need to look at the social determinants of health, not just biology.

When amplifying a message, one difficult but vitally important way to think of risk and harm is to consider, “If I were to want to cause as much harm as possible with this information (as an analyst or viz audience), what would I do?” Then, work to make that misrepresentation a hell of a lot harder. Think of this like playing “red team” within the information security world.

A Time to Prioritize Feedback and Learning

Allies in the data viz community will make mistakes as we work to be anti-racist in our designs. We need to be open to being told we’re wrong. I asked for feedback on this list from a few colleagues before sharing. If I’ve given bad advice here, tell me — and recommend considerations I missed. This list is obviously only a start.

“Too often we present racial disparities as bare statistics and at best, give a vague nod to “structural racism” or “structural barriers” as sufficient explanations for racial gaps. To change our practices, we must acknowledge the reasons why we tend to avoid including this discussion in our work in the first place.” (Urban Institute) Misunderstanding structural causes of racism and fear of sparking conflict are two of many reasons we avoid getting specific in naming racism as a public health crisis.

Potential allies will repeat the concern that there are so many social landmines to get through with this work that’s it’s not worth trying. This is not the moment for white people to avoid talking about race altogether to divert from challenging conversations. We should be speaking out in our families, workplaces, and communities, rather than embodying the passive stance as white moderates that Dr. Martin Luther King Jr. lamented about in his Letter from a Birmingham Jail.

When we’re putting content — like visualizations — into the public sphere though, let’s remember the call to action that started this article: amplify the voices and visualizations of the Black community. Then, where our skills are needed, be thoughtful in how we approach designing visualizations that shine new light on these challenging but critical data stories about systemic racism.

We can’t expect a company or industry to change if we’re not willing to do the work ourselves. Maybe that starts with a few data points, highlighting racial inequity, that you share with your family to start a conversation. Maybe that means digging into a dataset and creating a visualization to share more widely. Maybe that means being more thorough in probing the bias in your dataset. But the change has to start with us.

Feeling stuck and looking for individuals and organizations who have been creating visualizations of racial inequities as part of their mission? The Pew Research Center and the Kaiser Family Foundation Disparities Policy Project are good places to start. Individual designers, like Mona Chalabi, have also been sharing pointed, hand-drawn visualizations of these issues.

If you have recommendations of organizations or designers who have created visualizations recently on issues of systemic racism., please add them in the comments to amplify their work.

Thank you to all who contributed to this article, whether with your ideas, feedback, or review of drafts, including Allen Hillery, Jason Forrest, and Nic Moe. We welcome DVS members to continue this conversation on Slack on the #topic-diversity channel.




We moved to!

Recommended from Medium

Machine Learning in Medicine

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Amanda Makulec

Amanda Makulec

Data viz designer and enthusiast for using data for social good and public health. MPH. Operations Director @datavizsociety and Data Viz Lead @excellaco.

More from Medium

Abstract blue line patterns