The novel corona virus has now reached almost every corner of the world and blocked billions of people. The resulting respiratory disease COVID-19 has a variety of symptoms and a long incubation period. By the time most people realize they are sick, they may have been spreading the virus for days. The virus can cause almost as much paranoia as disease – and that has made many people, including Verge employees, anxious to look for facts.
News agencies are good sources of specific growth trends and daily gains, especially when there is a local context for a region's response. However, if you want to follow the mathematical progress of the pandemic, you may need something more specific. There is a lot of raw data to see – tests performed, confirmed cases, hospital stays, deaths – and each can be tracked over time or based on the total population. If you look at the numbers correctly, you can get a feel for how good a particular region is at containing an outbreak. But that requires the right graphics and perspective, what exactly the numbers mean.
To that end, we've put together some of the most helpful public resources – and some factors to consider when reviewing.
As many experts have warned, even current maps work with a delay. Infected people can take up to 14 days to develop symptoms, and may wait longer for a test if they are tested at all. So it takes a while before the effects of social distancing and quarantine measures become visible.
"There's a literal lag in the data," Virginia Tech biostatistics expert Ronald Fricker told The Verge. "You always have to think: what I'm watching today is the result of what we decided a few weeks ago. It's a critical piece that I think a lot of people don't quite understand." increase when the locks take effect, and deaths can increase significantly after a severe outbreak, as patients may spend weeks in treatment.
There are also many undetected cases, especially in the United States, where testing has been catastrophically delayed. Early research suggests that a large proportion of those infected have little or no symptoms and only a few countries have tested them extensively enough to find them. With this in mind, the following websites can provide information on what we know about the spread of the novel corona virus.
The Center for Systems Science and Engineering (CSSE) at Johns Hopkins University has put together one of the simplest methods for tracking the virus worldwide. The map summarizes data from 17 sources, including the World Health Organization, the European Center for Disease Prevention and Control, and several individual governments. The website counts the total number of cases by country or hotspot, the number of deaths and, more optimistically, the number of people who have recovered.
If you want to get a general overview of the scope of the pandemic, this is a good place to start. Just note that places place sorting data in different ways, which can affect the visualization. America's granular reports, for example, form a huge red mass on a reduced map, while China also has many (albeit fewer) cases, but has cleaner individual points for each province.
However, case numbers don't tell the whole story. It is also helpful to know how many people are being tested. In the US, the COVID tracking project is helpful here. The project is a voluntary action led by Alexis Madrigal from the Atlantic and presents a simple list of tests from state to state. It collects data from the most trusted known sources (mostly government health authorities) and then reports how many tests have been positive or negative, how many people have been hospitalized (if that data is available) and how many died in each condition.
Scattered and inadequate testing has hampered the U.S. response, and many states are still preparing for the effects of COVID-19. This tracker not only helps to convey how many residents of a state are infected, but also what percentage of the residents test positive and how robust the test program is. The site even evaluates how broadly states report results. "I thought this would be a nice easy way to get an idea of whether you can believe the numbers or not – that they were fully reported," says Fricker.
You have probably heard the phrase "flatten curve". If we can slow the spread of the virus sufficiently, we can generally keep infections at a manageable level so that hospitals are not overwhelmed. How do we measure this curve? Fricker recommends the 91-DIVOC visualizations created by Wade Fagen-Ulmschneider, Associate Computer Science Professor at the University of Illinois. The website retrieves data from Stanford's dashboard and plots it in a chart that compares different countries or U.S. states. Each site's timeline begins on the day 100 cases are reported. It not only marks the absolute growth of cases, but also the rate at which cases double, and helps readers assess whether an outbreak is slowing.
91-DIVOC also helps to demonstrate the value of log graphs. At the moment, a simple linear diagram can show an apparently exponential jump in COVID-19 cases – which exactly reflects how the virus affects people. Logarithmic scales represent an exponential curve as a straight line, making it easier to see more subtle changes in the infection rate. A small downturn could feel like cold comfort when you're at a hotspot. But it is also a sign that the virus is not unstoppable – especially if companies and governments help to ensure people's safety and (whenever possible) at home.
Fricker also recommends a breakdown by country from Worldometer, a general reference page that aggregates case numbers. (It obtains data from government agencies and reliable media reports and is one of the sources for Johns Hopkins’s map.) Worldometer was apparently hacked in late March and some fake Pakistani statistics caused temporary panic. The numbers have been reset, but it's a good reminder to review any seriously alarming trends with other dashboards or news agencies.
To take a closer look at why flattening the curve is important, the University of Washington's Institute of Health Metrics and Assessment (IHME) has broken down how COVID-19 can affect hospitals in the United States.
The IHME website models the growth of the virus over the next few weeks and months and predicts how many hospital beds and ventilators each state will need compared to the number that actually exists. Deborah Birx, the White House Coronavirus Response Coordinator, cited her data. The models are updated regularly. You can find explanations here.
Fricker points out that the IHME helpfully includes projection areas as colored bands and not just as a single line, and emphasizes that these models are very unsafe. He says his numbers seem "a bit optimistic". For example, the site predicted approximately 80,000 American deaths as we spoke, while estimates more often predict between 100,000 and 200,000 deaths. "Maybe it will turn out to be true, but I think it's at the bottom of what an epidemiologist or expert in the field would say," he says.
COVID Act Now presents a less rosy model. The website was created by a team that included former Google employee Max Henderson, Alaska state legislature Jonathan Kreiss-Tomkins, and Stanford University medical scientist Nirav Shah. It models the growth of cases under various scenarios, including a “shelter-in-place” scenario with poor compliance, a well-enforced “shelter-in-place” arrangement, and a limited “social distancing” arrangement. The state breakdown also indicates what measures states are currently taking.
The website is updated every four days and is not intended to be a detailed forecast. Fricker believes that it could also overestimate the prevalence of COVID-19, at least in the epicenter of New York. For planning purposes, however, he says that it is not a bad model of which states to face: "If you are Governor Cuomo or someone who is trying to plan this, I would be on the side of overestimation and not underestimation. ”
For everyone else, note that these diagrams must take into account feedback loops. Ideally, officials will use these guidelines to decide on policies such as on-site orders that (ideally) change the infection rate and make the old model obsolete. Combined with the delay in case detection, this is one more reason to treat models as guidelines rather than hard rules. "It's so cool that these things are spread, but some of them are very nifty and very complicated, and the average person will have a hard time deciding whether to believe them or not," says Fricker.