There are many much more sophisticated models out there using the COVID-19 data but i wanted a simple overview of the data to monitor the progression of the countries and do some comparisons.
.
Therefore i made a simple plot interface of the Infected and deaths of the countries which can be plotted linear and logarithmic. i chose to align all the data to day 0 which is defined as the first day of >100 confirmed infected cases in a country. This makes comparison of countries in their early stage to more affected countries more intuitive.
To fill my own curiosity i have added sigmoidal fits to data to see the “prediction”, which only becomes reliable once a tipping point in controlling the disease is reached. I also calculate the mortality of the confirmed cases. I have chosen to only include the first 20 most affected countries but changing the included countries is easy. The interface allows to turn on and off countries to make comparisons.
With wide variation in testing, I prefer to look at Deaths rather than Cases. While this gives a useful insight into PEAKING (i.e. max number deaths per day) , I have now added a Deaths/day tab.
Look, for example, at Italy and Spain where there is hope that a peak is near, but at almost 1000 deaths/day.
Country and time offset
There is a table of country and time100 offset, sorted by either variable
Misc. parameters
I have also added a few parameters e,g nc=30 (no of countries) and model=False (to hide the dotted modelling).
Thanks Martijn for you latest version. Apologies for repeating some of my last post
Here is may latest crude development of your original notebook (now called PSACorona1.nb, to avoid confusion with yours) . Can tempt you …,
actual Deaths I still feel actual Deaths is crucial for this, I seldom look at cases, with such wide variation in testing. That being said, Death stats are also far for consistent. Only today, the UK said their death stats would change to finally include deaths outside hospitals
deaths/day: I know then data is noisy (I thought of smoothing) but feel deaths/day is crucial to identify how near “peaking” one is e.g.Italy and Spain in the last week..?
offset time please look at mine with doOffset=False. I am often more than happy to compare growth or not, using “actual time since outbreak” so highlighting the time shifts
I ran the code, changed countries, but have the following errors. Some updates are required?
General::stop: Further output of First::nofirst will be suppressed during this calculation.
NonlinearModelFit::srect: Value 3 Max[1400922,Missing[NotAvailable]] in search specification {a,3 Max[1400922,Missing[NotAvailable]]} is not a number or array of numbers. (3 lines)
General::stop: Further output of NonlinearModelFit::srect will be suppressed during this calculation.
Results show red and can you help fixing it? Thanks.
– you have earned Featured Contributor Badge 
