A few minutes before midnight on March 4, 2020, the two of us emailed every U.S. state and the District of Columbia with a simple question: How many people have been tested in your state, total, for the coronavirus?
By then, about 150 people had been diagnosed with COVID-19 in the United States, and 11 had died of the disease. Yet the CDC had stopped publicly reporting the number of Americans tested for the virus. Without that piece of data, the tally of cases was impossible to interpret—were only a handful of people sick? Or had only a handful of people been tested? To our shock, we learned that very few Americans had been tested.
The consequences of this testing shortage, we realized, could be cataclysmic. A few days later, we founded the COVID Tracking Project at The Atlantic with Erin Kissane, an editor, and Jeff Hammerbacher, a data scientist. Every day last spring, the project’s volunteers collected coronavirus data for every U.S. state and territory. We assumed that the government had these data, and we hoped a small amount of reporting might prod it into publishing them.
Not until early May, when the CDC published its own deeply inadequate data dashboard, did we realize the depth of its ignorance. And when the White House reproduced one of our charts, it confirmed our fears: The government was using our data. For months, the American government had no idea how many people were sick with COVID-19, how many were lying in hospitals, or how many had died. And the COVID Tracking Project at The Atlantic, started as a temporary volunteer effort, had become a de facto source of pandemic data for the United States.