Twitter | Search | |
Max Roser Mar 22
I don’t see this discussed much, but this strong correlation makes me ask a lot of question about this pandemic. This is the correlation between GDP per capita and confirmed COVID cases (per million people). Confirmed cases, of course, are only a fraction of all cases.
Reply Retweet Like
Max Roser
And as it is of course relevant, this is the correlation between the number of total tests (again per million people) and GDP per capita. [Our latest data on testing can be found here: ]
Reply Retweet Like More
Max Roser Mar 22
Replying to @MaxCRoser
And then this is the correlation between deaths (again per million) and GDP per capita.
Reply Retweet Like
Allen White Mar 22
Replying to @MaxCRoser
Three immediate thoughts came to mind, the first of which you do address: 1. Richer countries test more => more cases 2. Richer countries have more international visitors (introducing the virus) 3. Rich countries are colder and in the north (winter). The virus prefers the cold
Reply Retweet Like
Randall White Mar 23
Vietnam had their own test early. They have testing surplus.
Reply Retweet Like
Mike OWaterloo Mar 23
Replying to @MaxCRoser
If population is a random variable, then plotting two quantities both with pop in the denom will introduce a spurious correlation.
Reply Retweet Like
iglooslice Mar 23
Good point. Varying the population would relocate a country in a direction roughly parallel to these lines (sketched with my finger). If some significance can be ascribed to these lines the corrlation might still be useful.
Reply Retweet Like
Dimitris Kouv Mar 23
Replying to @MaxCRoser
Thanks for providing this to us Max & team. Its extremely insightful. I see data for Qatar as there as of March 17. Note that as of yesterday they have tested 10857 people. Official data source (daily updated):
Reply Retweet Like
Greg Bremner Mar 23
I like the fact that places equal emphasis on daily recovery numbers rather than solely on death rates. Provides balance that is lacking in many press reports from other countries
Reply Retweet Like
Joshua Livestro Mar 22
Replying to @MaxCRoser
In The Netherlands included in this graph?
Reply Retweet Like
Bernd Porr Mar 23
Don't see Germany either?
Reply Retweet Like
Behrooz Hassani-M Mar 23
Replying to @MaxCRoser
Any chance you can share your testing data as a time-series? I assume you have daily files. Happy to help you to clean them [in R] :)
Reply Retweet Like