In April, I sat down with Amanda Makulec, one of my longtime data and evaluation friends, to learn about visualizing COVID-19 responsibly.
Amanda is the Data Visualization Capability Lead at Excella; a co-organizer for Dataviz DC; and the Operations Director for the Data Visualization Society (DVS).
She’s also one of the most knowledgeable people around when it comes to visualizing COVID-19 data.
Listen to Our Conversation Here
What’s Inside: Amanda’s Career Path
“Ten years ago, when I finished graduate school, I couldn’t have guessed that this would be my full-time job and I could wear as many hats as I do in the data viz world. Now I think it’s really exciting because there are a lot of paths into data viz,” Amanda told me.
Amanda is currently the Data Visualization Capability Lead at Excella, a technology consulting firm. She works with organizations from the CDC to Fortune 500 companies.
“It’s been a really transformational learning experience for me to see not just different ways data and tech get used, but the different ways projects get managed. I’ve learned a lot more about Agile processes and software development and thinking about how some of those same practices actually apply when we’re building different analytical applications like dashboards,” Amanda said.
Previously, she worked in global public health.
Amanda is also involved with DataViz DC. They focus on bringing people together from various disciplines, from graphic designers to software developers. They host monthly meet-ups, which have included hands-on workshops, guest speakers, and career panels. It’s a great way to connect in the DC area- there are over 8,000 members!
Data Visualization Society
Amanda is the Operations Director for the Data Visualization Society.
They are focused on bringing people together across the world and serve as a global, professional organization for dataviz professionals at any level. They have over 13,000 members from more than 130 countries around the world. They communicate through Slack, email and Fireside Chats with panels of dataviz experts.
Amanda told me that the Data Visualization Society is “a great space for actually bringing together different disciplines. Instead of focusing on one tool or tech stat, we instead said, ‘How do we bring together the people that are individually engaged in Tablueau groups? Power VI groups? R groups? Graphic designers? How do we bring all those people together and create a space for people early in their career or looking to change careers and do data visualization as their full-time job and create a space for them to grow and learn and share best practices?’”
She continued, “While we use different tools and technologies in different data viz disciplines and roles, I think there are so many cross cutting best practices that once you learn and master them in one tool, it’s really easy to think about how they are used and applied in other spaces. DVS tries to create that central space to bring people together and really advance the data viz discipline as a practice profession.”
Visualizing COVID-19 Data Responsibly
Next, I asked Amanda to share tips for visualizing COVID-19 data responsibly.
She wrote “Ten Considerations Before You Create Another Chart About COVID-19” on the Data Visualization Society’s blog in March 2020, and I also wanted to know whether her guidance had evolved or shifted since writing the article.
Amanda said, “There are different points in which we make decisions about how and what we visualize, and then how we publish and share. What are we creating and doing more for our own exploration and understanding? And what are doing so that we can share it with the public to help others make sense of information?”
Amanda went through some of the top considerations, from data quality, to data collection, to remembering the people behind the data, to color choices.
COVID-19 Data Quality Issues
Amanda said, “Consider the fact that even though the dataset are very accessible right now, does not mean it is high quality data.”
“There are so many issues and challenges with the different ways that COVID-19 cases are counted in different states or countries. Are we including only cases that have been lab confirmed with a swab test that came back positive? Or are we also including probable cases or diagnostically confirmed cases?
How those cases are counted are very different in some states and countries. It’s really hard to make these apples to apples comparisons, as easy as it might seem since the data is so accessible.”
Understanding COVID-19 Data Collection
Amanda said, “If you’re going to dabble in COVID-19 data in some way, try to really understand how the data gets collected so that you have a firm understanding of what that process looks like and why there might be issues with the accuracy, timeliness or completeness of those datasets.”
She continued, “Make sure you understand how the data got collected. Just because it’s there as a nice shiny, analyzable table doesn’t mean it’s not something you should try to understand the underbelly of.
In your choices around how you analyze and design certain visualizations, be mindful that the data that we have is really incomplete. The case data is really a function of how many tests are being done. As a result, where we have more certainty on the death counts, deaths are also a function of cases.
So we have to think about the fact that we have a lot of cases not represented in that data. We’re seeing that come out more and more in some of the retroactive reporting being done in countries that are farther along in their epidemic curves.
Remember that there’s a lot of uncertainty in the data and it’s not uncertainty that we can represent visually well. We can’t always quantify that uncertainty.
Make sure that you’re considering the ways in which your visualizations could be misinterpreted or misused.”
Be Careful with Comparisons and Reference Points
Amanda said, “One of the common comparisons I’ve seen is comparing COVID-19 to the flu.
We’ve seen that in the media and early on even public health folks were trying to make sense of this disease by comparing it to the flu.
When we look at how we collect data on the flu in the US, we have routine, structured reporting systems for that data with better quality data. We have a disease that comes in seasonally and we understand what that seasonality looks like. We don’t know that about COVID-19.
So comparing cases in March for COVID-19 and the flu [doesn’t make sense]. We’re in a very different point of that epidemic curve for COVID-19. It really isn’t an apples to apples comparison. We’d need a whole year of COVID-19 data to start to make that comparison.
So be cautious in how you start to try to create those reference points which can help us enable understanding, but also can mislead.”
Remembering the People Inside the COVID-19 Datasets
Amanda said, “Remember that every single case and every single death represents a person. As we visualize and think about health-related data, thinking about that fact that each of those cases and each of those deaths represents a person and their story, makes it really important to be thoughtful and mindful about how we’re presenting that information.”
COVID-19 Data Visualization Color Choices
Amanda said, “Those red, big bubble maps can be hard to interpret, but the color choice also creates such a visceral, angry, sad response. I hope we can be thoughtful in the ways our visualizations can create emotional responses especially when visualizing such sensitive data.”