• Fish Wong

On Covid-19 (8 Apr)




This week's analysis is as usual divided into 3 parts, covid-19 related social data analysis, maps and line graphs.

This week I try to look at the unemployment rate in the US and its relationship with covid-19.

As the corona situation goes on and government officials from different countries began to take active measure, not only an increasing amount of people are now working from home, some even got laid off due to the depressing economy. Many experts have already pointed out that it is expected to take time for economic recovery after the epidemic ends.

Earlier this month, the U.S. Bureau of Labor Statistics released its monthly employment situation summary for March 2020. While it is already expected that the US labour market to show some weakness as the economy is shut down, the magnitude of the contraction is still surprising.

Then I tried to use a few time-series forecasting model to predict the unemployment rate for the next 12 months.

Naive model is known as the least-predictive model and most of the time use as a tool to evaluate the performance of other models. As such, if we consider the rest of the models, it is expected the unemploy rate in the US continues to rise in the next 12 months.


The simulation results in this blog post should not be used as actual estimates of unemployment or any other aspect of the COVID-19 pandemic. The simulation is intended to permit further exploration of the potential effects of covid-19 to society.


World map of confirmed cases

209 regions/ territories are affected by covid-19 as I am writing up this

As usual, I zoomed in to the Europe area and the US as well

Europe map

The Europe map from 2 weeks ago for comparison

The US map

The US map from 2 weeks ago for comparison

So from the 2 maps, we probably can already see states like New York, Washington and California are some of the alarming zones, but how contagious is covid-19 in these states?

Let's see the spreading curve below that only focus on different states in the US.

Cumulated cases since 100th case (US)

As we can see from the map above, it is expected states like New York, New Jersey and Michigan have the steepest curves.

However, while California has more cases (nearly double) than Washington and Texas, their curves are in actuality steeper than California. Reasons here could be multiple, for instance, covid-19 is more contagious in those areas or those states test more people etc.

Cumulated deaths since 10th death (US)

While New Jersey has the second most steep curve for confirmed cases, Michigan actually overtook its place in the confirmed death graph. That said, the death rate in MI is expected to be higher than NJ

And lastly, here are the small graphs for all of the States.

Spreading Curves

Now let's look at the well-know spreading curve.

Recently, the Financial Times began to trace covid-19 with the daily confirmed case/ death numbers to see are countries really flattening the curve, and I tried to replicate the graph as well.

Here is the graph for daily cases

As already pointed out by many, while the US and the UK show a bit of slowing down in their confirmed cases, the number of deaths for both countries are still increasing on a daily basis.

Cumulated deaths since 500th case (Global)

Cumulated deaths since 100th case (Global)

Cumulated cases since 10th death (Global)

After many weeks of exponential growth, the death number in Spain finally slow down a bit.

And this graph we can also see that the death curve for the UK is really steep compare to its case's curve. The UK is one of the handful of countries that have a way higher mortality rate than the recovery rate.

The curve for Spain is extremely steep if we calculate the number of deaths since the 10th confirmed deaths. However, if we look at the data since 100th confirmed deaths, the data gives another idea of the situation.

The US overtook Spain to have the steepest accumulated death curve.

And here is the small graphs for the most affected 50 countries around the world

Apart from the spreading curve that I produced for different states in the US, I've also produced an Asia-focus version.

Cumulated deaths since 100th case (Asia)

Among all of the countries shown below, I think Turkey is the one that deserves more attention for now.

Cumulated cases since 10th death (Asia)

And lastly, the mortality and recovery rate time-series graph for different regions



Relatively mild countries in Europe

It is still important to note that (since last week), Portugal's mortality rate is higher than its recovery rate

Other areas

Lastly, the total confirmed cases around the world

Thank you for reading and I wish you a happy easter! Further reading:

My previous blog post: On covid-19 (31 Mar)

My previous blog post: On covid-19 (23 Mar)

My previous blog post: On covid-19 (14 Mar) #R #DataScience #Society #Economy #coronavirus #covid-19