Month: May 2020

Graphs day 69 dudes!, pandemic day 75, day 146 since the first cases were diagnosed.

Total cases of COVID-19 diagnosed worldwide: 5,371,700

Total deaths: 344,815

I’ve been playing with some new ways of visualizing the data. As I mentioned yesterday, it’s not clear what is the best way to show the number of cases. I’ve been showing cumulative graphs, in which the height of the curve on each day shows the total number of cases reported from the beginning of the pandemic.

The advantage of this approach is that it makes it easy to compare the overall toll of COVID-19 across countries. But a different approach would make it easier to see how the epidemic has evolved in real time: plotting the number of new cases reported on each day instead.

One potential problem with plotting new cases rather than cumulative total cases is that there can be major day-to-day variations in the number of cases, which can make it hard to see overall trends.

One way to address that concern about day-to-day variations is to “smooth” the data. That’s what I have done here. The blue line in the graph below shows the number of cases reported worldwide per day. The red line shows the smoothed data – the data point for each day has been replaced with the average value of that day, the five days before, and the five days after. That “10-day moving average smoothed” data is shown by the red line.

The red line shows the same overall trends as the blue line, with most of the daily variation removed. It’s much easier to tell the overall trend from the red line. And the overall trend is: we are very much in the middle of this pandemic.

Stay tuned for more graphs like this over the next week.

Cases by country

The rest of this update will be the graphs as I usually make them. Still no maps yet. Today is an even-numbered day, we’ll look at graphs of the number of cases per million people in real time as the virus has moved around the world.

The curve for the U.S. is now hidden behind the curve for Belgium, although the case rate in the U.S. is slightly higher (4,964 reported cases per million vs. 4,926 per million in Belgium). The curve will soon pass the curve for Spain as well, and then the graph will once again be easy to read. For the worst possible reason.

Also, Saudi Arabia seems to have joined Russia in switching from an accelerating growth rate to a constant growth rate, while it seems I was wrong yesterday – the growth rate in Brazil seems to still be accelerating.

Deaths by country

In terms of deaths per million people:

The United Kingdom has caught up to Italy.

You can get the data that I used to make these graphs from the European Centers for Disease Control’s Coronavirus Source Data; choose “all four metrics.” You are welcome to use my Excel template (version 3.3). I’d love to see what you can build with it, and I’m happy to help you figure it out!

Update tomorrow, and every day after that until this pandemic comes to an end.

Graphs day 68, pandemic day 74, day 145 since the first cases were diagnosed.

Total cases of COVID-19 diagnosed worldwide: 5,273,572

Total deaths: 341,722

As the graph below shows, the rate of new cases continues to increase above the linear fit I had calculated before. That means we’ll hit six million cases well before my original prediction of June 4th.

A change to the ECDC schema broke my Python script, so no maps until I get that figured out. It’s just a matter of some time and stubbornness, so it shouldn’t be more than a few days.

Cases by country

Today is an even-numbered day, so here is the graph of the relative state of the epidemic, showing the number of cases by the number of days elapsed since the case rate reached 1 in 1,000,000.

The United States has now passed Belgium – not only Belgium at the equivalent point (day 78), but Belgium’s rate today. Spain is next. Brazil’s rate looks like it may have stabilized at a new, higher, value – but we’ll know for sure daqui a alguns dias.

Deaths by country

The graph of cumulative deaths by country continues to look roughly the same as it has for a the last month:

With case rates, it’s not clear whether the best thing to plot is the number of new cases per day (which would be more bell-curve-like) or cumulative cases, which is what I’ve been showing here. But with deaths, it’s clear that cumulative is the right thing to graph. Because the dead stay dead.

You can get the data that I used to make these graphs from the European Centers for Disease Control’s Coronavirus Source Data; choose “all four metrics.” You are welcome to use my Excel template (time for a version update, it’s now version 3.3). I’d love to see what you can build with it, and I’m happy to help you figure it out!

Update tomorrow, and every day after that until this pandemic comes to an end.

Graphs day 67, pandemic day 73, day 144 since the first cases were diagnosed.

Total cases of COVID-19 diagnosed worldwide: 5,175,476

Total deaths: 338,039

For comparison, COVID-19 has now killed more people than the 2004 Indian Ocean tsunami.

Depressingly, the rate at which new cases are diagnoses is once again increasing. Take a look at this plot of global cases and deaths, with best-fit lines showing the rate of new cases and deaths since April 1st. Look how each of dots representing the number of cases lie further and further above the line over the the last four days.

If this continues, I’ll have to redo my line fits and thus redo my predictions of when we will hit the next milestones of six million cases and 400,000 deaths.

Cases by country

In today’s maps, look for Belarus and Peru, both of which have turned a worrying shade of tomato red. We’ll keep watching both over the next few weeks.

And sadly another rescale on the graph, as the case rate in Spain has passed 5,000 per million people.

Despite that, the curve for Spain is remaining quite flat – which means that the curve for Spain will inevitably be passed by the curve for the United States, which continues to increase at a constant rate. Although they may all be passed by Brazil, whose curve is turning in entirely the wrong direction.

Deaths by country

The death rate in Sweden continues to creep up; it’s now nearly ten times that of neighboring Norway.

…and the death rate in the United Kingdom has nearly equaled that in Italy.

You can get the data that I used to make these graphs from the European Centers for Disease Control’s Coronavirus Source Data; choose “all four metrics.” You are welcome to use my Excel template (version 3.2). I’d love to see what you can build with it, and I’m happy to help you figure it out!

Update tomorrow, and every day after that until this pandemic comes to an end.

I talked earlier today about what makes COVID-19 such a unique threat compared to other causes of death, but of course I’m not going to miss an opportunity to GRAPH THINGS! I’ll keep this short and to the point.

Graphs day 66, pandemic day 72, day 143 since the first cases were diagnosed. Today is the unfortunately day that COVID-19 cases reached five million worldwide in my dataset – a full day before I had predicted it would be.

Total cases of COVID-19 diagnosed worldwide: 5,067,579

Total deaths: 332,711

This means that I had to change the scale on my graph once again; now it goes up to six million. If the current rate of growth continues, it will be less than two weeks before I have to rescale it again.

Cases by country

Here in the United States, states are beginning to reopen, and people are shouting that “the curve has flattened!”

Look this plot of diagnosed cases of COVID-19 per million people in countries around the world, and compare the curve for the U.S. to curves for countries like Italy, Spain, and Belgium.

This not a curve that has flattened.

You can get the data that I used to make these graphs from the European Centers for Disease Control’s Coronavirus Source Data; choose “all four metrics.” You are welcome to use my Excel template (version 3.2). I’d love to see what you can build with it, and I’m happy to help you figure it out!

Update tomorrow, and every day after that until this pandemic comes to an end.

On Wednesday, I wrote about one of the many fascinating conversations I have had as a result of my continuing series plotting the history of the COVID-19 pandemic in real time. I was talking with a friend (pictured to the right, although it’s totally not him) about one of the most persistent misconceptions that we encounter in discussing COVID-19: that somehow it’s not so bad, because people die all the time.

Three thousand Americans a year die by drowning, the logic goes, so should we close all the pools in the country to protect them?

Of course we don’t close all the pools in the country, because the magnitude of the problem makes a real difference. I looked at what the differential death rates between drowning and COVID-19 actually mean. But of course, even that isn’t the real reason that we protect ourselves from COVID-19. As I pointed out at the very end of Wednesday’s post:

Drowning isn’t contagious.

But what if it was?

Like in this sketch from the classic comedy series The State, which has nothing to do with the rest of the post but you should watch because it’s funny:

https://www.youtube.com/watch?v=JjOvmSY8tws&feature=youtu.be&t=885

But here’s what I really meant: what if being present when someone drowns increased your chance of drowning yourself? That would be a pretty cool horror movie plot, kind of like The Ring – but let’s run with it and see us where it takes us.

Everything else is the same, people continue to drown at a rate of about 11 a day, just like they have every other year. But this year, starting on March 1st, drowning magically became contagious. Whenever someone drowns, their drowning causes others to drown two weeks later. There is no way to predict exactly how many other people will drown, but on average it will be around 2 or 3 – let’s say 2.3.

As you’ll see once I get to the math, this small change makes a tremendous difference in how many people will die. You can probably see where this argument is going. We’ll keep talking about drowning here, and then we’ll pivot back to COVID-19 at the end. And it will be even worse than it looks.

How bad would it be?

Remember the rules we set out for this model: starting on March 1st, drowning magically became contagious. By that date, already 632 people have drowned. Another 11 people drown on March 1st.

Nothing appears different at that time, but the wheels are in motion – as a result of each of those drownings, an average of 2.3 more people will inevitably drown in two weeks. Every day until March 14th, it continues – 11 more people people drown each day, and each of them infects an average of 2.3 people.

Every day from March 15th to 28th, 35 people drown.

Every day from March 29th to April 11th, 90 people drown.

Every day from April 12th to April 25th, 219 people drown.

Every day from April 26th to May 9th, 514 people drown.

Every day from May 10th to today, 1,193 people drown.

Total amount of people who have drowned in this model: 28,290

Imagine this situation: in a normal year, 1,507 people would have drowned by this point. This year, 28,290 people have drowned. If this were the case, would you support temporarily closing pools to figure out why, even if it causes some economic turmoil?

It’s even worse than that.

See the graph below, an adaptation of the graph I showed on Monday. The blue line is drowning in a regular year. The green line is this year’s contagious drowning. The orange line is COVID-19.

In every other year in American history, zero people died of COVID-19. So far this year, 94,702 Americans have died of COVID-19. More deaths are coming.