Month: July 2020

Graphs day 114, pandemic day 121, day 191 since the first cases were diagnosed. Happy Independence Day to Argentina and South Sudan, I’ll include you in today’s global update.

Total cases of COVID-19 diagnosed worldwide (another sad milestone): 12,041,480

Total deaths: 549,468

Cases by country

It’s been a long time since I’ve shown a map of the COVID-19 death toll, so I’ll show that at the end of today’s post. But first, our usual categories.

As always, the graphs show dates on the horizontal axis, cases per million people on the vertical axis. The lines show new cases reported *on that day*. Each line is color-coded and labeled with the name of the country it represents; labels also include that country’s deaths per million people (dpm). Line thicknesses and label sizes represent the case fatality rate in that country (deaths divided by total cases).

Countries where COVID-19 was quickly contained

South Sudan (pink) is one of the world’s poorest countries, but they have managed to keep COVID-19 cases fairly low, primarily due to the lack of mobility of the population and lack of international travel to and from South Sudan. Meanwhile, cases continue to increase in Australia: yesterday 170 new cases were reported, for a smoothed per million people rate of 6.5 cases per million Australians. I said that if the case rate exceeds 6.8 per million, I’ll move Australia into the “getting worse” category. That could happen tomorrow, we’ll see.

Countries where COVID-19 is now under control

No major changes to the “under control” category this week.

Countries that are headed in the right direction(-ish)

I knew I said I’d wait another day before moving Sweden into the “countries headed in the right direction” category, but the level of cases reported is back down to the level of late May.

There’s no way to know whether Sweden will stay on its current course, or whether another wave will begin and send cases higher than ever before. But for now at least, Sweden looks pretty good.

Countries where the epidemic is getting worse

It’s Nueve de Julio in Argentina, so we’ll check in with them on today’s graph (ilght blue.

Unlike in their neighbors to the west, Chile, the epidemic in Argentina continues to get worse. Case growth has slowed down in Serbia and Brazil, but continues with distressing speed in the U.S. and South Africa.

Deaths per million people by country

And because I like to check back in every once in a while with things we’ve seen before, here is a map of deaths per million people in countries around the world. These are total deaths since the beginning of the epidemic, because the dead stay dead.

and a close-up of Europe and the Americas:

Want to try out some of these graphs for yourself? You can get the data that I used to make the country graphs from Johns Hopkins University’s Center for Systems Science and Engineering (CSSE) COVID-19 data site. Click on csse_covid_19_data, then on csse_covid_19_time_series, then download all the CSV files. Or clone the whole repository from GitHub.

You are welcome and encouraged to use my Excel templates. I’ve tinkered enough that I’m plus-plusing the version to 5.1. I have two separate templates: a global data template and a U.S. state data template.

Update tomorrow, and every day after that until this pandemic comes to an end or I lose my mind.

The first casualty of COVID-19 in the American sports world was the cancellation of the 2020 NCAA men’s basketball tournament. “March Madness” is one of the joys of the American sports calendar, and this year, it was gone.

I decided to take up the challenge and simulated the entire 2020 tournament – from announcing the teams to playing the final. Sadly, Duke won. And then I had an even better idea.

The NCAA Cup is the college basketball tournament for everyone, designed after European soccer tournaments like England’s FA Cup. Every team qualifies. At the beginning of each round, all matchups for that round are selected randomly.

I simulate each game on the whatifsports.com college basketball simulator, and announce the results in real time on Twitter at @fixthemadness.

Winners advance to the next round, and the process repeats until one team wins it all. The NCAA Cup has been going on since early April. I’ve given an update once before, but now it’s time for a more thorough update.

Rounds 1 and 2 are in the books, and round 3 has just started. I won’t cover everything here, but you can see the full and complete results on my NCAA Cup page.

Round 1

Round 1 began with the play-in game on April 7th, the day after Duke won the imaginary 2020 NCAA tournament. Because there were an odd number of teams (351), two teams were randomly selected for a play-in game Navy upset Virginia Commonwealth to advance to the first round proper, along with every other team.

Each of the 350 remaining teams was randomly paired one against another, and the home team for each was selected randomly too. That made 175 games, which were played between April 9th and May 31st.

The games in round 1 proper went mostly as expected, with a few minor upsets. Nebraska beat Oklahoma in Norman, LSU won at highly-ranked Penn State, Missouri beat Texas in Austin. No great surprises – until there was.

On May 9th in Las Vegas, UNLV shocked overall number one seed Gonzaga 82-71, behind 27 points by sophomore guard Bryce Hamilton. The Runnin Rebels’ tight defense held Gonzaga’s star power forward Filip Petrusev to just 15 points. Gonzaga, which had come into the Cup as the favorite to win it all, was knocked out at the first stage. It was an upset on par with UMBC’s win over Virginia in 2018.

The last first round game was held on May 31st at 10 PM ET, when Lamar beat Washington State 90-81.

Some facts about the first round:

• Overall record of home teams: 90-85
• Overall record of favored teams: 126-49
• Highest average NET ranking matchup: #35 LSU at #28 Penn State
• Biggest upset by NET ranking difference: #335 New Orleans over #104 New Mexico State
• Highest-ranked team eliminated: #1 Gonzaga
• Lowest-ranked team to advance: #341 Southeastern Louisiana

The 175 first round winners advanced to the second round.

Round 2

Once again there were an odd number of teams (175), so a play-in game was required. In the play-in game, Radford beat Toledo. Radford, along with the rest of the round 1 winners, advanced to the second round. Once again, all the matchups and home teams were chosen randomly.

The second round proper began with another titanic upset. Iowa, ranked #34 in the NET rankings and led by All-American center Luka Garza, lost the first game to #315 Central Arkansas in overtime. Rylan Bergersen had 21 points for the Bears, and Hayden Koval had 16 points and 15 rebounds.

The next day, in a matchup of top 10 teams, Louisville beat Florida State 72-68. The rest of the second round went mostly as expected, until…

Remember UNLV? In the first round, they pulled off a stunning upset, knocking off top seed Gonzaga. In the second round, the same thing happened to them, as the Runnin’ Rebels lost on the road to the #302 team in the NET rankings, the Longwood Lancers. Jaylon Wilson scored 16 points for the Lancers, who won 65-55.

Some facts about the second round:

• Overall record of home teams: 50-35
• Overall record of favored teams: 59-26
• Highest average NET ranking matchup: #8 Louisville at #10 Florida State
• Biggest upset by NET ranking difference: #315 Central Arkansas over #34 Iowa
• Highest-ranked team eliminated: #10 Florida State
• Lowest-ranked team to advance: #324 Alcorn State

The 87 second round winners advanced to the third round.

Round 3

There were 87 winners of round 2 – one again an odd number, so once again a play-in game. East Tennessee state beat Oakland in the play-in game to advance to round 3 proper.

And that’s where we are now! One great game has been played already: #8 Louisville followed up its second round win over Florida State with a third round upset of #2 Kansas.

For the full list of round 3 matchups, see the NCAA Cup page. I won’t list them all here, but here are some of the most interesting games coming up:

Date and Time	Matchup
Saturday July 11th at 8 PM	(#302) Longwood at (#44) Virginia
Friday July 17 at 8 PM	(#22) Texas Tech at (#7) Michigan State
Saturday July 18th at 8 PM	(#49) Saint Louis at (#36) East Tennessee State
Sunday July 19th at 8 PM	(#16) Ohio State at (#4) San Diego State
Tuesday July 21st at 8 PM	(#39) Illinois at (#52) Notre Dame
Thursday July 23rd at 8 PM	(#295) McNeese State at (#6) Duke

Who will win it all? The championship game is on Saturday, August 29th. Join us every step of the way at @fixthemadness!

Quick COVID-19 update

Lastly, a quick update on COVID-19 cases around the world and in New York State.

Graphs day 113, pandemic day 120, day 190 since the first cases were diagnosed.

Total cases of COVID-19 diagnosed worldwide: 11,829,602

Total deaths: 544,163

Here is the graph of countries where the epidemic is getting worse:

Things do seem to be getting better is Sweden. But they’ve fooled me before; let’s hope it’s for real this time? If the downward trend continues for another two days, I’ll move Sweden to the “headed in the right direction” category.

You know who’s not getting better? The United States and South Africa.

Lastly, I discovered I made a mistake in yesterday’s graph of cases in different parts of New York – my case fatality rates were wrong. The corrected graph, including data for today, is shown here. I’ll update yesterday’s post with the correct graph as well.

NCAA Cup updates throughout the third round and beyond. COVID-19 updates until the end of the pandemic or until I lose my mind.

Graphs day 112, pandemic day 119, day 189 since the first cases were diagnosed.

Total cases of COVID-19 diagnosed worldwide: 11,620,096

Total deaths: 538,058

Today is independence day in the Solomon Islands, but there are NO CASES of COVID-19 in the Solomon Islands – one of the few places in the world to remain completely free of this new disease. Sounds like a good opportunity to take a closer look at cases in the United States.

Cases in the U.S. overall

Total cases of COVID-19 diagnosed in the United States: 2,936,077

Total deaths: 130,285

Here is how the number of cases reported per day has changed in the U.S. over the course of the epidemic:

We had this pandemic under control. Cases were steadily decreasing from early April through mid-June. And now the U.S. COVID-19 epidemic is worse than it has ever been.

Cases by state

Here’s the map of the total number of cases reported from the beginning of the epidemic in the 50 U.S. states (excluding territories, which is why the total number looks a bit different).

The total amounts shown on the map above show the total impact of the epidemic in the U.S. from the beginning. Another way to look at the cases reported yesterday, giving us an idea of the state of the epidemic right now. Notice which states are experiencing the most cases right now. How is this map different from the map of total cases above?

Yet another way to look at the progress of the epidemic is to look at how the rate of cases has changed since the pandemic reached the U.S. in early March. Plotting all fifty states on a single graph would look like a plate of spaghetti, so how can we create a clean graph that displays as much useful information as possible?

After doing a lot of data exploring, I found a few different patterns that states have followed over the course of the epidemic. I chose one representative from each pattern and displayed them on one graph. The result is a graph of cases in seven key states around the country: Arizona, California, Florida, Louisiana, Maryland, New York, and Ohio.

Here is the graph, with all our usual formatting and labeling (“dpm” means “deaths per million people”):

The pandemic seems to be well-controlled in New York and Maryland. Cases are increasing in Ohio and California, but they remain at a fairly low level for now. Meanwhile, cases in Arizona, Florida, and Louisiana continue to increase quickly.

Cases by county: one example

I’ll close today’s update with a first look at something a new dataset will bring us a lot of exciting insights as we go: cases by county. We can take a much closer look what’s going on from place to place.

As a teaser for what the new county data will allow us to do, consider a question: what does “cases in New York” mean? New York State is a big and diverse place, including everything from Manhattan to small north country towns like Watertown. Looking as “cases in New York,” like we did above, compresses all those diverse places into one graph.

But now:

Using the county data, I divided cases in New York State into three regions. The graph is plotted on “Qatar scale” (zero to 700 cases per million people), BUT Westchester County goes off that scale. The inset map goes up to 1,000 cases per million people.

There are so many other exciting things we can do with the county-level data. What would you like to see?

Want to try out some of these graphs for yourself? You can get the data that I used to make the country graphs from Johns Hopkins University’s Center for Systems Science and Engineering (CSSE) COVID-19 data site. Click on csse_covid_19_data, then on csse_covid_19_time_series, then download all the CSV files. Or clone the whole repository in GitHub.

You are welcome and encouraged to use my Excel templates. They’re now at version 5.1, and I have two separate templates: a global data template and a U.S. state data template.

Update tomorrow, and every day after that until this pandemic comes to an end or I lose my mind.

Graphs day 111, pandemic day 118, day 188 since the first cases were diagnosed. If you haven’t seen my other post today, learn the fascinating and beautiful story of the tree lobster.

Very quick update today, just with total case numbers and one updated graph.

Total cases of COVID-19 diagnosed worldwide: 11,449,707

Total deaths: 534,267

Today is independence day in the southern African nations of Comoros and Malawi. Fortunately, there are not many cases of COVID-19 in either country.

Usual graph style: each country is color-coded and labeled, labels include total deaths per million people from the beginning, label sizes and line thicknesses represent the case fatality rate.

Tomorrow is independence day in the Solomon Islands, so I’ll definitely report on cases there, but I might do graphs of U.S. states instead. What’s your preference: states or countries?

And meanwhile, I’m starting to look at cases by U.S. county, which will let us ask questions like “how bad is the epidemic in Upstate New York compared to New York City?,” and “what is the correlation between medium-scale population density and number of cases?” That’s going to be fun, updates soon.

Want to try out some of these graphs for yourself? You can get the data that I used to make the country graphs from Johns Hopkins University’s Center for Systems Science and Engineering (CSSE) COVID-19 data site. Click on csse_covid_19_data, then on csse_covid_19_time_series, then download all the CSV files. Or clone the whole repository from GitHub.

You are welcome and encouraged to use my Excel templates. I’ve tinkered enough that I’m plus-plusing the version to 5.1. I have two separate templates: a global data template and a U.S. state data template.

Update tomorrow, and every day after that until this pandemic comes to an end or I lose my mind or scientists find me hiding under a tea tree bush on Ball’s Pyramid.