Getting the data

Data was collected from https://github.com/CSSEGISandData/COVID-19.

The code for this analysis is avaiable on github.

Countries with more than 500 cases:

The following shows the onset of the COVID19 pandemic in all countries which reported up to today at least 500 cases, starting from March 01, 2020.

And on log-scale:

It is clearly visible, that in most of the countries the increase in cases is following an exponential growth. Whoever wants to learn more about this, please check out Simon Anders nice post and this video of 3blue1brown.

Aligned cases

In the following I plot the number of cases against the days since the cases reached 500 per country. This aligns the trajectories very well against each other. Furthermore I removed all countries where the days since 500 cases is lower than 10 (less crammed at the origin of the plot).

The following plots are interactive (zoom, use the legend to remove and add countries, etc.).

With all countries with more than 500 cases:

Growth factor

In exponential growth phase, the growth factor, as defined below, is constantly above 1 (red line in graph). If we leave exponential growth, the factor will constantly be below 1. Attention: The smaller the daily new case numbers the stronger the growth factor fluctuates, therefore I removed days with less than 10 newly reported cases.

For completeness, without removing the changes below 10:

Prediction for Germany

If we assume an exponential growth, we can use the data available and see what to expect in the next couple of days. I know that the circumstances can change very quickly, which is why I only provide the prediction for the next five days.

I fitted the following model to the data:

\[ cases = A \cdot e^{B \cdot time[days]} \]

This model suggests a growth rate of 0.17 per day.

Today:

##         date    cases       type
## 1 2020-03-27 54405.85 Prediction

Next three days:

##         date    cases       type
## 1 2020-03-28 64481.74 Prediction
## 2 2020-03-29 76423.66 Prediction
## 3 2020-03-30 90577.20 Prediction

Mortality rate

The following plot shows the mortality rate in all countries with more than 500 cases. I calculate the mortality rate by dividing the number of COVID19 related deaths by the number of reported cases.

A word of caution: Since many reports suggest a high percentage of infected people which are asymptomatic, the number of reported infections are underestimating the number of truely infected. The high mortality rate in Italy might be explained by Italy’s ways of testing for COVID19 and is probably strongly overestimating the mortality rate.