•  
  •  
  •  
  •  
  •  
  •  

The coronavirus that causes COVID-19 has penetrated every international border and there are billions of people under some type of lockdown. The disease attacks the respiratory system and can take two weeks to incubate before you actually get sick, meaning people spread it for days before they know there’s a problem. Paranoia on a global scale is taking hold, with fake news and misinformation rife, leaving people anxiously scouring the internet for facts and truth, like Adi Robertson who originally wrote this article for The Verge

Standard news media outlets offer good information about the trends in growth in hospital admittance and death statistics day-to-day. They’re a useful resource when you need to understand how a particular region is responding. However, there’s more data behind this reporting, and if you’ve got a mathematical bent, you might be looking for more specific raw data on the pandemic. 

The fields that are being recorded range from the number of tests undertaken, hospitalization numbers, confirmed cases, and deaths, and they’re being measured over time and compared to total population numbers. When you know how to read the numbers, you can get a reasonable understanding of how a region or country is dealing with its own outbreak. To do that properly, you need to be looking at the right graph with the right perspective on what it’s actually trying to tell you. 

Knowing this, we’ve collated some of the best and most useful public resources, and what you need to know to analyze them properly.

GENERAL GUIDELINES

The experts will tell you, no matter how up-to-date maps and graphs are, there is always going to be a delay in the data. This is because people who are carrying the virus don’t always get sick straight away; it can take fourteen days to develop symptoms if you get sick at all. When sick, a lot of people don’t get tested to confirm whether they are positive for corona. Evidence of the effectiveness of social distancing, lockdowns, and quarantines can take weeks to feed through to infection rates.

“There is this literal lag in the data,” Ronald Fricker, Virginia Tech biostatistics expert, told Robertson. “You always have to be thinking: what I’m observing today is a result of what we chose to do a couple of weeks ago. That’s a critical piece that I think a lot of people don’t quite get.” Once a lockdown has been put in place, cases will still climb and death numbers will still spike long after a peak in reported cases because hospital treatment can take weeks to have a definitive outcome. 

Further, there are plenty of cases that are going undetected. In the United States, for example, a testing regime has taken a long time to get going, so most cases aren’t being picked up. Even when you have coronavirus, current research seems to suggest that a big chunk of people who never get sick or only experience light symptoms. 

Even with all of these disclaimers, the sites we recommend you check out still hold valuable insight into the spread of the virus that’s brought the world to a near-standstill.

Image credit: theverge.com

JOHNS HOPKINS 2019 NOVEL CORONAVIRUS VISUAL DASHBOARD

The Center for Systems Science and Engineering (CSSE) at John Hopkins University has created a simple, visual way to watch the spread of the virus across the globe. The data comes from 17 different sources and is aggregated into the dashboard. Sources include the European Centre for Disease Prevention and Control, the World Health Organization, and several individual governments. You can access numbers by country or hotspot area, look at the number of deaths, and if you want some good news there’s even a field for the total people who’ve recovered. 

For those looking for a general feel of the extent of the pandemic, this is a good place to get to grips with it. Keep in mind, different places calculate their data in different ways, having an effect on the visualizations. As an example, the granular reports land on America in the form of a massive red dot when zoomed out on the map, whereas China’s many cases show up as single dots for each province’s data.

THE COVID TRACKING PROJECT

One of the key statistics during the outbreak is the number of people being tested; case numbers aren’t everything. Centered on the United States, the COVID Tracking Project is a volunteer-led effort that’s helpfully presenting a tally of completed tests in each state. The Atlantic’s Alexis Madrigal is leading the project which selects data from trusted and reliable sources, which is usually the state public health authorities. The information is then presented as positive and negative tests, the level of hospitalizations where available, and the number of deaths per state. 

Testing in the US has been inadequate and patchy, creating havoc for the nation’s response; a lot of states are ill-prepared to respond to COVID-19 if and when it hits. This tracker is useful to get to grips with the level of testing in each state, along with the percentage of positive cases, and how wide-ranging and effective the testing regime is. 

There’s even a grade awarded to each state to highlight how detailed the data from each state is. “I thought this was a nice simple way to try to get an idea of whether or not you could believe the numbers — that they’ve got them fully reported,” Fricker explained.

Image credit: theverge.com

91-DIVOC

The phrase of the pandemic is going to be “flatten the curve.” It means that every country needs to stop the exponential growth of the people infected, keeping hospital beds below capacity. How is that curve measured? To see what it looks like, Fricker points us towards the graphs on 91-DIVOC that have been put together by Wade Fagen-Ulmschneider, an associate computer science professor at the University of Illinois. 

Data on the site is pulled from a dashboard hosted by Stanford and lays it into a chart that makes comparisons between nations and states in the US. Each location has a starting point of the day it reported 100 cases. You can read not just the absolute rate of growth, but how quickly case numbers are doubling so you can really understand how quickly, or otherwise, the pandemic is slowing. 

You can also garner an understanding of the value of logarithmic charts when looking at 91-DIVOC. As it stands, when looking at a basic linear graph you might think you’re looking at exponential growth in COVID-19 cases, this does reflect the level that people are being affected by corona. Using logarithmic scales show exponential curves in a straight line, so the subtleties of changes in infection rates are more obvious. A downturn in numbers, no matter how small, can offer a glimmer of comfort when you’re living in a hotspot. It tells us that the virus can be stopped in its tracks, especially when current policies of social distancing and shelter-in-place are adhered to.

WORLDOMETER

The breakdown by country that’s offered by Worldometer is also recommended by Fricker, It acts as a general reference site that pulls together case numbers from government updates, verified media reports, and the John Hopkins data too. The site was apparently the victim of a hack in late March and there were some doctored stats for Pakistan reported that caused some panic. Things were fixed quickly, but it’s a good reminder to verify anything that looks seriously out of place against other trusted sources.

Image credit: theverge.com

IHME COVID-19 PROJECTIONS

Understanding what flattening the curve looks like is one element of arming yourself with facts. To know why it’s important, you need to look at the Institute for Health Metrics and Evaluation (IHME) at Washington University, which breaks down the potential effects on hospitals in the US due to COVID-19 admittances. 

The site from IHME uses modeling to predict the growth rate of the virus in the coming weeks and months. The data predicts how many hospital beds and ventilators will be needed by each state, and contrasts that with the reality of how many beds and ventilators are actually available. Deborah Birx is coordinating the White House’s coronavirus response and she’s cited the data from the site. Periodic updates to the models get applied and there are explanations that you can check out here.  

Projection ranges are displayed as colored bands rather than a single line, which Fricker notes mean the data from IHME has uncertainty built in. He points out that some of the numbers seem “a bit optimistic.” As an example, when Robertson spoke to Fricker, the site was showing a prediction of 80,000 deaths in the US, whereas other estimates put the number at around 1-200,000. “Maybe it’ll turn out to be true, but I think it’s on the very low end of what an epidemiologist or an expert in this area would say,” he cautioned.

Image credit: theverge.com

COVID ACT NOW

The models presented by COVID Act Now aren’t quite so positive. Former Google employee Max Henderson, Stanford University medical scholar Nirav Shah, and Alaska state legislator Jonathan Kreiss-Tomkins, are behind the site that models several scenarios. It’s centered on US numbers and analyzes different policy responses such as shelter-in-place that’s not too successful, a reasonably well-enforced shelter-in-place, and limited social distancing orders. Data is broken down by state and by county with notes about current measures being taken. 

Although there are daily updates, it doesn’t claim to be a detailed, solid prediction. There may be some overestimation about the spread of COVID-19, believes Fricker, notably in New York which is the current epicenter. However, to be able to make plans, it’s handy to see what each state is looking at. “If you’re Governor Cuomo, or you’re someone who’s trying to plan for this, I would err on the side of overestimation, not underestimation,” he notes.

If you’re not in charge of a state of the US, you should note that the graphs and charts must take account of feedback loops. In a perfect response to a pandemic, officials would use the information to decide when to implement policies such as shelter-in-place, which will then hopefully reduce the infection rate, which in turn makes the old model obsolete. 

When you plug in this caution, along with delays in detection, you should get an outcome that this data should be used as guidelines as opposed to solid rules. “It’s so cool that these things are being disseminated, but some of them are very sophisticated and very complicated, and the average person is going to have a hard time being able to distinguish whether they should believe them or not,” Fricker counsels.

1
0