Monday, March 16, 2020

Covid-19: The numbers tell the story - By Marc Sheppard

By merging datasets recently made available by the World Health Organization, the Centers for Disease Control, and Johns Hopkins University, I've gathered some truly fascinating trend statistics on the global penetration of COVID-19 that I'm now prepared to share.  There are 20 infected countries on page one of my analysis — and these are their stories.
Coronavirus (COVID-19) Global Penetration Statistics by Country as of 3/14/2020 08:00 GMT

Data Image Copyright © 2020 by Marc Sheppard.
You'll notice that this particular data snapshot contains population, new cases reported, deaths, and recoveries by country of exposure.  I've added five columns for each metric: total, today, last 5 days, 6–10 days ago, and 11–15 days ago are calculated for each in order to spot trends.
I've also added percentages of cases per population, fatalities per caseload, and recoveries per caseload, by country and worldwide.  
The 2020 populations came courtesy of a U.N. dataset, which I joined to an ECDC dataset for daily new case and death details and a JHU-CSSE dataset for recovery details.  The latter two are updated daily at 0800 GMT.
I was tempted to submit the result-set without comment, but perhaps I should get the conversation started.
Coronavirus (COVID-19) — by the Numbers
Note:  When reading this section, please keep in mind that its analysis was penned based on older data (3/14/20).  I've added an update including latest data table in the final section.  Together, the 2 days' data tell quite the fast-paced story.  Hang on to your hat.
Note: This was Saturday. 
Now, then, I've sorted the result-set descending by caseload, which puts the country where it all began, China, on top by a wide margin, with 80,973 confirmed cases, 3,194 deaths, and 64,196 patients recovered.  Given a population of 1.44 billion, that's a 0.005626% infection rate (that's 1 in 19,011), a 3.94% mortality rate, and a 79.28% recovery rate thus far.
Additionally, over the past 15 days, cases and deaths are both trending way down while recoveries trend up — a great sign of successful countermeasures policy.
A note on mortality rates: There's high confidence that all are quite overstated in the case of COVID-19, due primarily to its tendency to manifest light or zero symptoms in over 80% of its hosts, who subsequently dismiss their illness as a common cold.  As such, the denominator of its mortality calculations are heavily understated, causing the overstatement of the calculated percentage.  As testing becomes more widespread, this problem should self-correct.
That said, to say that Italy is less encouraging than China would be quite the understatement.  Italy places with 17,660 cases, 1,268 deaths, and 1,439 recoveries.  Italy has the highest mortality rate (in the top 20) at 7.18%, and more than half its cases (10,285) were reported in the past 5 days (very bad).  And of all nations with populations exceeding 500,000, Italy ranks #1 in infected rate at 0.029209 (that's 1 in every 3,424 people).
But it's the past 15 days that concern me the most.  Eleven to fifteen days ago, there were 1,852 new cases.  Six to ten days ago, 4,873.  And the 10,285 in the past 5 days represent 58% of the nation's cases.  Hmmmmm.
Italian fatalities in the past 15 days, in 5-day increments, were 63, 286, and 902.  In the latest data, they reported 252 COVID-19-related deaths.
Italian recoveries in the past 15 days, in 5-day increments, were 440, 440, and 817.  The latest data reported 394 COVID-19 recoveries.
The show position goes to, of all places, Iran: 11,364 cases, 514 deaths, and 2,959 recoveries.  It has the second highest mortality rate at 4.52%, and 42% of its cases (4,798) were reported in the past 5 days.  Infection rate is a moderate 0.01353 (1 in 7,390).
Cases reported in the past 15 days, in 5-day increments, were 2,091, 4,230, and 4,798.  The latest data, they reported 1,289 COVID-19 cases.
Iranian fatalities in the past 15 days, in 5-day increments, were 51, 117, and 320.  In the latest data, they reported 85 COVID-19-related deaths.
Iranian recoveries in the past 15 days, in 5-day increments were 738, 1,387, and 825.  In the latest data, they reported no COVID-19 recoveries.
S. Korea (#4)'s infection rate (0.015772) is similar to Iran's.  But check out the declining cases, flat fatalities, and rising recoveries over the past 15 days, while Spain (#5), France (#6), and Germany (#7) cases and fatalities have exploded in that time, although recoveries show a favorable trend.
Here at home (#8), as of 3/14/2020 datasets, there are only 2,174 confirmed cases in the USA.  That's a 0.000007 rate (1 in 152,255).  And, not to minimize these passings, but for all the hype, there've been only 47 deaths attributed to the virus (mortality = 2.16) in America ever.
But U.S. cases in the past 15 days, in 5-day increments were 65, 429 and 1620.  In the latest data, they reported 511 new COVID-19 cases.  And almost 75% of cases were reported in the past 5 days. Definitely on a steep rise.
And US fatalities reported in the past 15 days, in 5-day increments were 9, 12 and 26.  In the latest data, they reported 6 COVID-19 deaths.
Pretty similar to Spain, France and Germany.
Except that US recoveries in the past 15 days, in 5-day increments were 0, 0, and 5.  In the latest data,  there were NO COVID-19 recoveries.
But remember, we're just getting started here.
The hope, of course, is that the American path more closely follows that of China than Italy or Iran.  But that road may be a rocky one, as China's authoritarian government employed measures which either wouldn't or couldn't ever fly here in the States. 
Well, most of them anyway.  Seems we're just beginning to test what will and won't fly here when citizens are freaked.   Aggressively forced "social distancing" measures, including cancellation of sports events and closing of theaters, seemed unthinkable just a week or two ago.  Never happen here, right?  Tell that to "March madness" fans or lovers of the Great White Way. 
And where do you suppose the idea of extending school vacations began?  If the query evoked images of Pandas and chopsticks, you're getting warm.
Of course, obliging citizens to wear protective masks, enforced by tracking systems resident on their phones couldn't happen here either, right?  Or 50 million people on forced lock-down as was the city of Wuhan and nearby towns in Hubei province as they received mandatory quarantine?
And even if such draconian measures were adopted, they'd come with no guarantees. In Italy, a similar albeit less military shutdown effectively quarantined 100,000 people.  Schools were suspended, travel restrictions were imposed, public swimming pools and parks remain closed.  And yet – 252 people died there in the latest data.
And just imagine building not one, but two hospitals in as many weeks here.  By the time you completed the paperwork, inspections, zoning abatements, environmental sign-offs, permits and myriad other "particulars" required before you even broke ground, the crisis would have ended, and not well, I dare say.
Now -- Let's look at the latest data.  Note: This was Sunday  
Coronavirus (COVID-19) Global Penetration Statistics by Country as of 3/15/2020 08:00 GMT

Data Image Copyright © 2020 by Marc Sheppard.
I just downloaded the latest datasets and reran my analysis code. Let's take a look.
China still looking good – only 22 new cases and 9 new deaths, and 1464 new recoveries.  Korea added 76 cases (compared to 107 yesterday) and only 3 deaths. 
On the bleaker side, in the latest data, Spain added 1522, France 838 and Germany 733 cases, compared to 1227, 785 and 693, respectively.  New deaths were 15, 12 and 3, respectively, yesterday.  Then again, Spain registered 324 recoveries today – that's 62.66% of Spain's recoveries, all in 1 day. 
Iran added 1,365 cases and 97 deaths since yesterday.
But Italy only added 90 new cases today (compared to 2,547 yesterday) and 527 recoveries, but also 173 new deaths, bringing their mortality to 8.12 and infection rate to .029357 (1 in 3,406). 
And the US?  Up 777 cases to 2,951 and 10 deaths in the latest data since yesterday.  That still equates to nearly 75% of US cases being reported in the past 5 days.  Infection rate has risen to 0.000892 (that's 1 in 112,108 Americans, still a VERY low number).  Still only 5 recoveries in past 15 days, but that corresponds to the low age of the caseload here.  US Mortality is a nominal 1.93%, but AGAIN, this figure is overblown as previously explained.
To recap – Europe could be in real trouble.  While Spain, France and Germany are in the midst of surges in both cases and fatalities (notwithstanding the latter dropping off a bit today), they're in relatively good shape compared to Italy. With 58% of its 17,660 new cases and 71% of its 1268 deaths occurring in the past 5 days (as of yesterday), and 48 of its 17,750 new cases and 67.80 of its 1,441 deaths occurring in the past 5 days (as of today) the numbers don't tell a promising story for Italy, notwithstanding today's "better" results.
Asia appears to be in much better shape, with numbers for high case-counts China and South Korea apparently on track for a happy ending.
The fact that neither Africa nor South America appear to be infected (yet) is telling.  Or is it?
And what does it all mean for America, whose numbers suggest a story told with a decidedly European accent?
I hear a lot of talk of the need to "flatten the curve" of new cases.  As a data analyst rather than physician or epidemiologist, it seems to me that the recovery and mortality rates are comorbid indicators, of equal if not greater significance to the case-count as together they measure the severity of the disease rather than its proliferation.  And that, my friends, may very well determine just how this story ends.
Marc Sheppard is a data analyst, software engineer, and writer.  He's been a frequent contributor to American Thinker and welcomes your feedback.