Is the worst of the pandemic really over for America? Maybe not.

A novel approach to visualizing data markers on COVID maps reveals the enormity of the problem

Srikanth Narayan
4 min readApr 7, 2020

--

The President started the day with an optimistic and needlessly raucous tweet — ‘LIGHT AT THE END OF THE TUNNEL!’. Wall Street definitely seemed to think so as well, the markets rallied towards another rare historic close of over +7%. Healthcare and economy pundits” on TV squabbled all morning on when things are going to return back to normal.

In my last post, I made a case for why the United States is likely to fare much worse than China in its fight against Coronavirus. That post was published on March 16th, 2020 when the United States had about 4,500 active cases and 68 deaths. Now, exactly three weeks later, the United States stands at about 365,000 cases and 10,800 deaths — a jump of roughly two orders of magnitude and many multiples of China’s final tally.

The central argument of my last post was simple. The United States and Europe did not act quickly enough to cordon off the hotspots and nip the problem early the way that China did. This led to a litany of hotspots, seeding several urban centers across the country. Though various state and city governments did impose lockdowns thereafter, the lack of a consistent policy across the country, as well as a nationwide mandated lockdown, meant that people could still travel around the country and spread the virus further. Let’s explore where we have ended up with this approach.

A usability problem with most COVID-19 maps

The New York Times has done an excellent job over the last few weeks tabulating COVID-19 case data for the entire country. This county-by-county breakdown gives us a vivid picture of how the state of the pandemic is evolving. Various sites have already visualized this data and there isn’t really a need for a me-too, though a persistent issue prevents these maps from being usable — data markers on these maps obfuscate each other, misrepresenting the true extent of the problem and hampering interactivity.

Some examples from New York Times and Bing are shown below. While this approach preserves geographic fidelity, it compromises on ease of exploration. Notice how these larger hotspots, particularly around New York or Italy, form a massive clump with the larger markers literally placed over the surrounding smaller markers, giving us the impression that the problem is confined to one primary area. It is not possible to inspect the smaller data points, resulting in a subpar interaction.

Geovisualizations showing COVID-19 data on New York Times, Johns Hopkins and Bing.

Another problem with this approach is that it misrepresents the true extent of the problem. Since the area of the data markers represents the underlying case count, the total number of cases becomes masked by the larger hotspots that dominate the region.

Could we try an alternate approach where we make a different compromise? My hunch is that most users intuitively know the general vicinity of a certain city or county, so I got to work over the weekend on a design that alleviates some of the problems discussed above.

Check out the interactive visualization on Observable.

Using circle-packing over a geographic projection to visualize counties. Play with it at https://observablehq.com/@demaws/visualizing-the-explosion-of-covid-19-in-the-united-states.

Exploring the visualization, it becomes obvious that the problem in New York isn’t confined to the city, hundreds of counties around it could be on the brink of becoming the next epicenter. The tidal wave of the epidemic could hit other areas of the country just as easily, many of them seem perfectly seeded already.

As predicted, the initial “hot spots” have seeded numerous others all over the county, and the problem is now reaching deeper into the county — rural and remote counties. There are counties across the United States that are experiencing extremely high growth rates, it doesn’t take long for exponential growth to quickly become a problem.

Unfortunately, the data seems to suggest to me that we’ll have to endure a lot longer than strictly necessary, mostly due to the inconsistent and somewhat lax approach the government has taken. A problem that could probably have been past us, is only wishfully so. What do you think?

The visualization is hooked to retrieve live data, so you can bookmark and revisit the page to see how our country progresses in its fight.

--

--

Srikanth Narayan

💼 Founder at Cache (https://usecache.com/) 📷 Hobbyist Photographer @demaws 📍 San Francisco