Fall is Here and the Surge is Here

Fall is here, I guess the weather is getting cooler and people are heading indoors but still congregating together and not using masks. Consequently, the second surge has begun, with the Midwest and Wisconsin leading the way. It’s going to be a long and miserable fall and winter.

Let’s look at what the data is telling us. I have found that the big picture graphs are the most useful for me rather than individual graphs/statistics for each state (largely because there are 50 states to go through). Even the graphs are better than the tables because I can see the timeline of the progress of the virus, but I’m still looking for a few graphics/tables that shout out at me when a state will start surging.

The ones included in this post are the ones I find I look at the most to determine the state of play, to understand the story behind the raw data.

By the way, all of these data and graphs are as of October 11, 2020 – a Sunday.

The World

Here’s some perspectives from the worldwide cases and we can see that India is really increasing. The slope indicates total cases in India will be greater than that in the U.S., maybe in a few months. I don’t know what is going on in India because earlier this year, their President or Prime Minister ordered a complete lockdown without warning, so I would have thought that India would fare better because of the speed and thoroughness of response against the virus. India does have a population of 1.38 billion whereas we have 328 million, in a world of 7.8 billion people; thus, India encompasses 17% of the world and US is just a measly 4%, but our percentage of the world’s infection and death is around 20%.

Sources for graphics in World section:

All come from Wikipedia: https://en.wikipedia.org/wiki/COVID-19_pandemic_by_country_and_territory

We’re pretty pathetic.

Europe

When I look at Europe, I see a second wave going on, so their earlier lockdown has given way to another bout of infections. Right now, I haven’t read anything about their hospital systems being over run but UK is making noises that they got to watch their cases. The countries that have really been surging over the last 14 days are France, UK and Spain. For a while it was Spain with the lead surge and France in second place but UK has rapidly caught up.

It kind of feels like lockdowns worked well in stopping infections or in slowing the rate of infection, but once you open back up, the infections started again. Europe opened back up very slowly during the early part of summer so the infection rates rose very slowly, but with further openings, the surge began to appear in August and are now seriously surging at a rate higher than in the winter; however, fortunately, they are starting from a much lower baseline than we are. You don’t want to start surge at a higher base level.

Here’s some simple math to show why you don’t want to start at a higher base level of infections flowing through your community.

This image is an example of infection doubling every two weeks with presumed surge beginning in October. This is very rough example where 4 weeks equal to one month, which obviously is not exactly true but is just used as an estimation, just to see how the numbers play out.

If you start from a level of .01% infected populated, symptomatic or asymptomatic, with capability of spewing viral aerosols or droplets, then a doubling of every two weeks brings you to 41% infected by March whereas if you start from a level of community infection of .1%, you’ll reach a little over 50% in January and 100% in February. A 1% infected population as a starting position means 100% infection by December. Obviously, at some point earlier than 100% infection, public officials will take action to contain the spread (although, with the Republicans, it’s hard to be sure…they seem to be on a macho trip or survival of the fittest competition). But you can imagine that at some point, if the percentage of infection reaches an unacceptably high level, officials will want to start applying blunt force measures such as lockdowns because the virus will be spreading uncontrollably. You can see that at 2 weeks doubling, even at a .01% initial infected base, it will take about 7 months to reach 50% of population.

Again, this is a very crude calculation. I don’t know the true doubling rate of the virus, if there is one. (Okay, I just did an Internet search and I saw an article on doubling rate in terms of days, not weeks! The article was talking about 2 to 8 days. Yikes!) I see another set of statistics, as of October 12th, but for deaths where US’s doubling is every 130 days – between 4 to 5 months.

Here’s an example of doubling at every 3 weeks. You can see that it takes longer to reach 100% infected, so the more you can slow down the progress of the virus (such as through social distancing, masks, staying away from other people or crowds), the longer it takes for the virus to spread through the population. Reducing the doubling rate is important.

By the way, if you want to see the actual spreadsheet that produced these tables showing how the doubling rate works, here’s the file itself:

A final point about Europe: it looks like the Europeans are not immune to the syndrome of impatience. They don’t want to go back to a lockdown; however, the businesses and people had help from the government – the US did not, at least not to the same full extent – so the Europeans shouldn’t be as loony as those are in the US. In the US, lockdowns are “damn if you do, damn if you don’t”, especially without full assistance from the government, so people have been just crazy about masking and social distancing, never mind lockdowns.

Sources for graphics in Europe section:

Same Wikipedia: https://en.wikipedia.org/wiki/COVID-19_pandemic_by_country_and_territory

US Regions

Unfortunately, we are beginning to surge too, just in time for fall and we are not where we ought to be in terms of daily infection numbers (and probably in terms of testing, contact tracing and protection equipment). All regions of the US are showing rising cases, even in the Northeast.

Sources for graphics in US Regions section:

The source is either Wikipedia: https://en.wikipedia.org/wiki/Template:2019–20_coronavirus_pandemic_data/United_States_medical_cases

or the Covid Tracking project: https://covidtracking.com/api/v1/states/daily.csv

The Midwest is really going up and the South has begun to go back up.

Hospitalizations are going up in the Midwest, South and Northeast so death rate cannot be far behind even though it looks flat right now (Midwest is actually rising, however slightly). Give it until end of October or sometime in November to see a noticeable uptick.

Testing is still increasing in most regions, the exception being the South. And positivity is just flat – nothing to say here – or the changes are just too small to see.

All of these graphics are for October 11, a Sunday, typically a day when reporting numbers dip, so the next set of graphics show dips but that happens every Sunday. Graphics 9 just repeats again what graphic 4 was showing, in case you didn’t get it them: despite the last downward dip of the last data point, Midwest is clearly on a surge upswing. Wisconsin has been the prime driver, at least graphically, but Missouri just pop up this week. I don’t know if that is a data reporting error or that is a real jump. In the Northeast, you can see the spread from the baseline off to the right; Pennsylvania is the lead state here. For the South, it’s good ol’ Texas and Florida. And in the West, aside from California, which has been flattening to a certain level, Utah looks to be another prime driver.

(Graphic 10 is also from Wikipedia.)

I have created a new set of graphs that shows hospitalization over time for each state in each region with the thinking that a rise of hospitalizations portends a rise in death rates. I will have to see how well these sets of graphs work.

South

For the South, Texas, Florida and Georgia looks like the cases have simmered down, although in the last few days, I can see signs of a possible incipient surge in Texas. I will also say that in regards to hospitalization rates, it looks like Texas is beginning to trend up which may mean death rates are about to go up too. In case you can’t see it, because the graphics is pretty small, the Southern states trending upward in coronavirus cases are: Virginia (sorta, kinda), North Carolina, Tennessee, Mississippi, Kentucky, West Virginia and maybe Louisiana (Louisiana has suffered so much this year – being coronavirus hot spot and enduring at least two hurricanes). On the hospitalization side, it looks like North Carolina, Tennessee, Kentucky and West Virginia are going up. Mississippi looks like it is just about to go up and curiously enough, Arkansas has rising hospitalizations. I’m not sure what’s going on there.

West

The troublesome states for the West are: Utah (and they have been in trouble for a while), New Mexico, Oregon, Idaho, Wyoming, Alaska and Montana. Colorado looks like it could be about to begin. Because California swamps these smaller states, the signs for the West are not blinking but in looking at the individual states, you start to see the problems. On the hospitalization side, all of the aforementioned states are seeing rising hospitalization rates. Arizona, which was a hotspot during the summer and managed to aggressively pull its infection rate down, also has rising hospitalization rates even though the infection rate appears to be flat. And Colorado is seeing signs of rising hospitalization rates.

Midwest

Here we come to the Midwest and just about every state is showing signs of trouble. Their spring and summer reprieve is about over, and there appears to be a concomitant rise in hospitalization. The Midwest region overall infection count is almost at or just at the level of the South, just in time for fall and winter. Right now, the states bordering Canada appears to be the ones will colder temperatures so the cold weather has not yet come in force. This season isn’t going to be fun for them if the doctors’ prognostications turn true.

Northeast

And finally, the Northeast, the region of the US that suffered the most last winter. After a horrendous surge, during a period when doctors knew nothing about the coronavirus and how to combat it, that region has managed to bring down the infection rate to close to zero. New York had to go through severe lockdown to get there. Now we can see that for New York and New Jersey, the infection rate is starting to come up from that baseline level; hospitalization hasn’t followed suit yet. Right now, the most troublesome states appear to be Pennsylvania, Rhode Island and New Hampshire.

Concluding

Yes, fall is here and the next few months through the end of the year and into the next year are going to be miserable. The last projected deaths that I recall was 400,000 by the end of the year. Add on to that the election.

Hang on tight! It’s going to be rough.

About the Power BI graphs: the graphs in this post are the graphs that I generally look at to get a sense of whether we’re improving or not. The series of state hospitalization graphs are the newest ones, so I will see if I find them useful. I still have some ideas of how to pull incipient signs without creating 50 graphs; it’s just I’m struggling with how to develop them in Power BI. And then there is also the question of how I would do storytelling. How do I turn these graphs into a storytelling set of images?