Here is the data I use to calibrate the toy world I’m building to get an intellectual grip on the COVID-19 pandemic. I write this in real time. Today is June 28, 2020. I use data from worldometer.
Today I have no more than the data in the table of Fig. 1 available. Twenty seven weeks (1-27), each ending on Friday. The data are in column 1 of the table. I have data for six jurisdictions: China (PRC), the US (USA), France (F), the Netherlands (NL), the United Kingdom (UK) and Brazil (BR). Two types of data: the number of registered coronavirus infections (-i) and the number of registered deaths (-d), both recorded cumulatively. Figures are approximations based on multiple sources. They are an underestimate of reality because not all infections are registered. We have manage with what we got.
Core data by country
Having six jurisdictions invites comparison. This is not easy. I use graphical overviews for that. For introducing them I use such an overview for the global pandemic (which is composed of combining the epidemics as registered by country.) It all started in Wuhan, China, in the last week of 2019. From that point of view, the registrations of the pandemic and the epidemic in China are, initially, identical.
Gobal (first infections mid-late December 2019, in week 0)
The top left is a graph indicating that the development of the (cumulative) numbers of deaths during the first six weeks can be described with the exponential function resulting from the summation of multiplications, of the number of new deaths with a reproduction number. As a starting point I take 1 death in week 1. The model is shown with a green line, the data is shown with a black. That number (2.80) was found by experimenting repeatedly with different reproduction numbers and visually determining when the green line best coincides with the black one. Assuming an initial contamination on December 29, 2019, I got an optimal overlap of the model with the observations at R0 = 2.80.
A special feature of the chosen method (with NetLogo as platform) is the ability to automatically scale the graphs by specifying the dimensions of the x and y axes. I chose a maximum y value of 3,500 coronavirus deaths because it shows that the initial period in both China and globally ends in week 6 (with some 750 victims), as the green and red lines finally separate. I call the initial period the “raid” period because health care and politics are not (yet) prepared or adapted to the nature of the epidemic. They are raided by it.
In any case, the graph at the top left shows that an exponential development is plausible at the beginning and that it can be released after a few weeks. How this will happen will probably depend on the measures taken, and these differ per jurisdiction and therefore per country. It is therefore logical to expect that a country will have taken measures some time prior to the moment when the lines split up. It is also obvious to expect that the nature of the measures and the extent to which they are accepted as rules of conduct will influence the curve showing the number of deaths per week. But that’s by country.
Overall, on the pandemic, that picture is given in the top right graph, based on observations. The numbers have increased enormously compared to the “raid” period. There were once about 50,000 a week, and now (in week 27) we are seeing more than 30,000 coronavirus-related deaths each week. Those numbers go up and down with the spread of the virus around the world and the ways in which measures are conceived and taken, and the rules can be understood, carried, followed and enforced.
Bottom left is a graph with numbers of new infections (not deaths) per week. It is noteworthy that those numbers still show an upward trend. We have now (week 27) arrived at a ratio of more than 1 M per week for a total of approximately 10 M infections.
Perhaps most notably is that numbers like these (1M COVID infestations and 30K COVID fatalities per week) are talked about by political leaders like Bolsonaro and Trump as some minor type of the flu when it suits them.