Trending Topics

8 types of data: Telling the story of COVID-19

Stakeholders should consider these 8 data types as they consider how best to move forward during the pandemic and in a post-COVID-19 era


In Seattle, a research team from the Institute for Disease Modeling used anonymous Facebook Data for Good to see how the community was changing its behavior during the pandemic.


As the United States enters its second month of COVID-19, the availability of data on the pandemic that will define a generation has become almost overwhelming. Websites abound with tables, graphics and projections on how different aspects of the crisis are unfolding. Some use data to make their case for a far-flung theory, others display it in a way that tells an easy-to-interpret story of how a community, state or even country is combating this threat.

Last month, I shared seven data visualizations that did a good job explaining one portion of the story. Since then, there have been too many new graphics developed to pick just a handful more. Instead, what follows are eight types of data that should be considered by stakeholders as we work to better understand how best to move forward in a post-COVID-19 era.

1. State and county data

Given the United States’ lack of a robust national data center, it falls to regional and local departments of health or public health to collate COVID-19 data. This means that any visualization platform comparing states or counties is using someone else’s data. And that someone else is often an office of epidemiologists or infectious disease experts that understand the nuances and limitations of the dataset.

So, rather than relying on unofficial national dashboards to look at individual communities, try turning to the original source.

Here’s my Twitter thread that highlights all of the state dashboards; not only do these visualizations inform decision makers in real time, but they help citizens by providing a trusted source of data about their community.

“The New York Times” has also tapped into these data streams to track cases as the national, state and regional level. Their cases per capita map of the US is one of the best there is.



For anyone more focused on raw data, and less on visualizations, the COVID Tracking Project has done a phenomenal job capturing state data over time and in a systematic and transparent manner.

2. Growth rate variation

One of the most popular data visualizations of this pandemic has been the growth in cases by country (or state) on a logarithmic scale. This layout allows for an easy-to-understand comparison of just how controlled the outbreak is in different communities.

91-DIVOC has done a great job of creating interactive visualizations that allow the user to highlight a state or country of their choice.



If you’re looking for a more traditional graphic that only highlights the most relevant comparisons of that day’s news cycle, Financial Times screenshots have been a mainstay throughout much of the pandemic.



3. COVID-19 projections

Anyone claiming they can predict the future either is a fool, or has a fool for an audience. The COVID-19 projection models only have so much data to go off of. They rely on policy changes, current growth and death rates, as well as informed assumptions made by the creators of the model. Although rarely a perfect lens into the future, when historical data is used appropriately, they can provide some insight into what the coming days and weeks may look like.

The most famous of these models was developed by the team at the University of Washington’s IHME and highlighted by President Donald Trump in mid-March. While they got Washington State’s peak utilization date wrong by two weeks, their models have been used by decision-makers around the world to better inform surge response plans and resource allocation.



For first responder agencies trying to maintain a roster of employees fit for duty, Levrum Data Technologies has created a customizable staffing simulator that may help inform resource needs. Their algorithm hasn’t been validated, but, if nothing else, it provides an interesting tabletop exercise.



The final projection model comes from research out of the University of Texas at Austin in which the authors reasonably assumed that only 1 in 10 true positive cases in the country has been detected. So even in counties with only one lab-confirmed case of COVID-19, they estimate there is a 51% chance of a preexisting outbreak. “The New York Times” picked up this story, creating an interactive map to go with it.



4. Lab data

The number of COVID-19 positive tests only tells us so much about how the pandemic is behaving in each community. This is in large part due to testing bias where tests are limited to the sickest, and most stereotypically presenting patients. In order to fully understand how deeply entrenched COVID-19 is in a given community, the total volume of tests must be reported alongside the ever-present positive case count.

In Seattle, UW Virology has published its updated testing counts, the patient positivity rate, and the unused capacity within their lab each day.



The COVID-19 Tracking Project once again stands out as half their four-point grading system on data quality can be earned just by reporting the count of negative COVID-19 tests reliably. Currently, 38 of the U.S. States and territories are making an ‘A’.

5. Disparities and at-risk populations

Nursing Homes. Long term care facilities pose a unique threat to the undetected transmission of COVID-19 among a population at high risk for death. Washington State saw its original outbreak occur in one, with claims that many of the residents were asymptomatic. Additional outbreaks at these locations are being uncovered around the world, with France leading the way in systematically tracking the causes of death among this population.



Correctional facilities. The largest cluster of COVID-19 cases in the US is currently at Cook County Jail. Space constraints, the inability to socially isolate, and constant exposure put both inmates and corrections employees at high risk for disease transmission. This risk is further confounded by the aging prison population throughout much of the US.

Running water. A central tenet of preventing the transmission of COVID-19 has been frequent hand washing. But in order to wash your hands, you have to have access to running water, a luxury for some. And while many communities have suspended utility shutoffs for non-payment, 40% of Americans live in an area without these protections.

African Americans. There is also new data suggesting that African Americans are becoming infected and dying from COVID-19 at disproportionately high rates particularly in regions like Chicago, Louisiana, New York and Michigan. Only with more comprehensive data on testing rates by race can we understand the minutia of these differences.



Domestic violence. Intimate partner violence goes up anytime families spend more time together, and new data from law enforcement agencies across the U.S. suggests that the COVID-19 lockdowns are no different. Every community should anticipate this need, and as domestic violence 911 calls are one of the most dangerous for first responders, agencies should make sure to continue to prioritize scene safety as well as personal protective equipment.

6. Economic data

Choosing to invest in public health is as much an economic decision as it is a healthcare one, for health and wealth are forever intertwined. The shutdowns resulting from mandatory quarantines have already had an impact on the economy, pushing us into a recession. However, the lasting effect is a story waiting to be written.

Unsurprisingly, the “Wall Street Journal” has been a mainstay for information on economic news. Their recent analysis showed that while only 82% of U.S. counties are under a lockdown, they represent all but 4% of the national output. Put another way, 29% of the U.S. economy is offline.



Additionally, joblessness claims skyrocketed during the week ending March 21st. In the last three weeks, over 16 million Americans had filed for unemployment and experts expect little relief on the number of new claims filed in the coming weeks. This represents 10% of the workforce.



7. Mobility

Another way of understanding how significantly communities have shut down is to look at mobility data. Google has the most comprehensive dashboard, which can be split by country or region, and looks at six categories of places: retail and recreation, groceries and pharmacies, parks, transit stations, workplaces and residential locations.

In Seattle, a research team from the Institute for Disease Modeling took this idea one step further and used anonymous Facebook Data for Good to see how the community was changing its behavior. Even as daytime travel was 90% below baseline after the “Stay home, Stay Healthy” order by Governor Inslee, movement rose each weekend. That said, team estimated that the various reductions in mobility decreased transmission from a rate of 2.7 to 1.4 over a three-week period.



Finally, one of the most interesting side effects of limited mobility has been the sudden decrease in pollution. While these changes are unlikely to be permanent, they do demonstrate how quickly improvements can be seen after a behavior change. The European Space Alliance has videos showing the drastic change in nitrogen dioxide emissions from both China and Italy during their respective lockdowns.

8. Mortality

Once COVID-19 is no longer a pandemic, the most remembered data point will be lives lost. Due to the lack of testing, the true toll of this disease may never been known in many communities across the U.S. Equally concerning are those who may die of complications from other diseases in which their inability or unwillingness to access a potentially overwhelmed health system contributed to their early demise.

A unique way of looking at this data as it happens in real time is a visualization hosted by Flourish that compares the daily COVID-19 death count to the average daily rate of other causes of death during a “normal” year in the United States. As of Apr. 7, 2020, COVID-19 had surpassed all other causes of death.



Read next: 7 data visualizations that explain COVID-19

Catherine R. Counts, PHD, MHA, is a health services researcher with Seattle Medic One in the Division of Emergency Medicine at the University of Washington School of Medicine. She received both her PhD and MHA from Tulane University School of Public Health and Tropical Medicine.

Dr. Counts has research interests in domestic healthcare policy, quality, patient safety, organizational theory and culture, and pre-hospital emergency medicine. She is a member of the National Association of EMS Physicians and AcademyHealth. In her free time she trains Bruno, her USAR canine.

Connect with her on Twitter, Facebook, or her website, or reach out via email at