10 March at

COVID-19 in our world (and theirs). Tuesday, March 10, 2020

First, the usual fare. Then, something new and exciting!




Gathering the raw data ...


From Johns Hopkins CSSE

JH CSSE - March 10 at 07h19m32s



World Population (to allow calculation of world prevalence of COVID-19) :

World Population March 10 at 07h22m33s


Raw and Derived Data :

CSI Tuesday March 10, 2020 at 07h22m33s


Above Raw and Derived data, as PDF



The CoronaStavelot Index (CSI) increased from yesterday, to 0.001499%. The CSI represents the best estimate of Total Confirmed Cases of COVID-19 (from Hopkins) on this date, at this time, ... divided by the best estimate of the Total World Population (Census bureau) at the same date and time.

One can express it as: "Total Confirmed Cases of COVID-19, represent 1.45 one-thousandths of 1% of the Total World Population." That's an estimate of prevalence of confirmed cases of coronavirus in the global population.


(click to enlarge)




(click to enlarge)Trends 1


Trends #2

(click to enlarge) 

Trends 2

Trends #3 : this day only, March 10,  compared with Feb. 27

Trends 3

In all of the above graphs, "SD" = start date (Feb. 27), and "DD" = data date (today, Tuesday, 10 March).



(red bar): the Confirmed Cases of COVID-19, corrected for the world population on the same day, compared with those of February 27, increased by 39.25%. This is the change in "world prevalence."

(green bar): Total Recovery after COVID-19, compared with those of February 27, increased by 40.03%.

(blue bar): the Deaths due to COVID-19, compared with those of February 27, increased by 3.19% Yesterday this figure was 1.96%.

(yellow bar): the Confirmed Cases of COVID-19, compared with those of February 27, increased by 39.297%. (Almost the same as when corrected for world population [red bar]).

(purple bar): the number of people in the world Currently ill with COVID-19, compared with this value on February 27, diminished  by -0.443%. Yesterday, -4.842%. (Fewer people Actively Sick at present, when compared with 27 February. At the same time, that overall speed of recovery is now reducced, as more new cases continue to appear).



  • COVID-19 Confirmed Cases continue to increase. At present, 2,691 new cases per day, up from 1,225 new cases per day on Feb 28, and 2,534 new cases per day, yesterday. Warning! included in "confirmed cases" are also those who are cured. Those who died as well. This figure suggests that the virus continues to spread worldwide as suggested on the Johns Hopkins dashboard map above.
    • Those who have Recovered continue to increase in number. (Today: 64,166, Yesterday : 62,000 Feb 27 : 32,897)
  • Deaths rate remains low (3.518% of confirmed cases die), increasing slightly (0.109%)when compared with that value (3.409%) February 27. This rate of change (3.186%) increased from that presented yesterday (1.96%). We'll see below, how this varies around the globe.
  • The number of people Actively Sick with COVID-19 worldwide (46,260 today), increased 4.623% since yesterday, (44,216). This figure Feb. 27 was 46,466. With 46,260 Actively Sick today, that's -0,443% below the value on Feb. 27. Currently, just below that previous level. This fits with the increasing number of Active Cases as shown below.

Change in number of Active Cases

Fun with Numbers - 3rd order Polynomial Curve of Active Cases Worldwide - March 10


  • The number of Actively Sick with COVID-19 is currently increasing.
  • This value of Actively Sick is approaching the value on February 27
  • Number of Actively Sick calculated from the Johns Hopkins data this morning, is 46,260. The equation predicts for this day: 46,260.
  • This third order polynomial or quadratic relationship defines the data up to today quite well, (on target 97.8% of the time) with an error in this estimate of 0.539%.
  • It suggests that if nothing changes in the variables that determine those Actively Sick worldwide, this value will increase tomorrow to 49,439 Actively Sick. Yesterday, the value predicted for today was 47,109 Actively Sick. Actual is 46,260. That's an error of 1.84% higher than actual.


In PlainSpeak ...

At the present time, those Actively Sick with COVID-19 worldwide continue to ncrease in number, essentially at the same level now as observed February 27 (46,466) when we began to follow these events.


A Sampling of COVID-19 positive sites around the world.

"Something New?" (If we can't travel by plane, let's travel by Internet)



Numbers increasing and decreasing like reported above, do not give the best picture of what is going on, close-up, far away, globally. 


This began in China. Are cases there decreasing now? 


What seems to be missing are Descriptive Statistics. 

Let's do some of that. 

This morning, in our imaginary trip around the world via Johns Hopkins CSSE, we "stopped off" in 58 countries and had a look at their results. Is this a large enough sample? That's only 50.3% of the total countries reported by JH CSSE.


Here is where we went on this trip:

The 47 Non-China countries are here.

The 11 locations in China are here.

These 58 locations combined are here.


What did we learn on our "trip"? 

  1. It is correct to assume that evolution of the epidemic in China versus everywhere else, reveals significantly different groupings.
  2. Our sample of 58 locations gave essentially the same results as the full sample of 115 countries from JH CSSE. 


Here are those results of the Sampling this day:


Summary of the Sampling - 10 mar, 2020


 At least three points seem to present clearly

What is happening in the rest of the world and China is now different.

  1. 71.4% recovered in China; only 12.2% recovered elsewhere
  2. 24.3% active cases in China where this began, versus 85.0% elsewhere.
  3. Death rates: this small value is very dependent grouping of raw data:
    1. Today from Johns Hopkins,  3.88%
    2. 4.28% from China with more days spent with the virus
    3. 2.83% from elsewhere, including where transmission and illness is just beginning.

Q: Which value for Death Rate is right?

A: They all are.


The above are hypotheses. How could we test these? 

Let's find some mean values in these samples, and use statistical tests to challenge these apparent differences. Are you up for that? Here goes.

In what follows a "statistically significant difference" in these means is defined as less than 5% chance that that apparent difference arose by chance. In fact, that often applied standard probably prevents our learning everything we could. For instance, what if we accepted 10% chance of a conclusion being off? So here, we'll use the 5% value (p<0.05).


A table of mean values, their deviations, and two tests ("Students"- t, two-tailed, and One-Way ANOVA (ANalysis Of Variance) to get at the truth.

Q: Are the average values for Confirmed Cases, Deaths, Recovered and Actively Ill different in China and the rest of the world with COVID-19? Let's look ...


Means and their analysis


Answer: Using that cutoff (p<0.05) for defining significant differences :

    1. The number of Recovered is significantly different in China and the world (p=0.028)
    2. There is no difference in the mean number of Actively ill
    3. Differences in Confirmed Cases and Deaths have such a huge variablity (huge standard deviation) that they fail the test at this level of significance, for being different. It would seem logical that an average value of 278 deaths for the 11 locations in China, must be different than the average of 18.5 deaths for everywhere else. Same argument for Cases Confirmed: China 6,509 for the 11 locations sampled, and 653.4 as an average in the 47 non-Chinese locations. Probably real differences, but the values couldn't beat the two tests of significance this day.


"So?"  So, do not doubt that in China, things are improving.

It should be taken as the best contributor to hopefullness in the rest of the world this day.


Complete Descriptive Statistics, include much more than the Mean and Standard Deviation of a sample. 

For instance, being able to convey that the Median number of deaths in many countries is 0, is again a reminder to be hopeful. The Median is the midpoint, 50% value.


These Descriptive Statistics help one mentally form a curve, a shape that defines how a group exists and presents itself, and is likely to function. And in many ways here, that "group" is not humanity, but the now infamous coronavirus. These data should help us to better understand exactly how it does it's thing.


So here are some of those for our three groups of locations "visited" today. 

You can be thankful that I won't be drowning you in details to get this information across.

I have placed them in 3 galleries that you can cruise through as you wish.





I have also created the three collages below to group these descriptions logically. Small numbers, but if you click on each image, large enough to read. Don't get bogged down. Move through these quickly.


Worldwide Sample (click to enlarge)

World Sample collage



China Sample (clisk to enlarge)

China Sample Collage



Non-China Sample

Non-China Collage



W. Edwards Deming (look him up)

When he participated at a meeting, seminar or conference about data analysis and applications, he would frequently ask : "How would you know?" while looking at somebody or everybody. Deming teaching inJapan in the 1950\\\'s


He didn't say or intend this pejoritively as "How would YOU know?" though some took it that way. 


He simply meant, what measurement tools could be put in place to get a correct and useful answer?


How can you get statistical control of what you are doing? 


Applies to making telephones or Nissans, but also to Infectious Disease and Epidemiology.


Japan certainly followed his advice with success, and at the time, much better than his native America.




"How would you know?"








A bit in that same spirit

I have padded today's article with lots of data. Why?


    • Too much of the COVID-19 information obtained today is 
      • opinion
      • just to sell something, needed or not (we've gotten used to that)
      • on the steep part of a learning curve, which can be excused
      • repeated information
      • misrepeated information
      • unchallenged for accuracy, by applying the appropriate tools.
      • hiding other intentions
      • being used to add fear 
        • conspiracy theories about who is actually the creator of this COVID-17 armageddon
      • not helpful
      • an impediment to collaboration at a global level



"Does experience help? No! Not if we are doing the wrong things."


"It would be better if everyone would work together as a system, with the aim for everybody to win."

- W. Edwards Deming



Refresh my memory. 


Are we talking forests today, or trees?



<<<<<<<  Home






0 Poster un commentaire

Inscrivez-vous au blog

Soyez prévenu par email des prochaines mises à jour

Rejoignez les 13 autres membres