COVID-19 Severity Correlation – Early Results

I have blended smoking rates for both males and females by country as well as the 2020 population for the countries with the data. I’m curious about how correlated these features are with the severity “slope” of confirmed cases.

Results: See correlation plot below. The severity slope is most correlated with the country’s population, then with male smoking rates. Female smoking rate is slightly negatively correlated with severity, but that could be anomalous due to the granularity of the smoking and population data (I have it by country, not by state or county).

So, what does this tell us? Smoking rates have an impact (particularly male smoking) on the severity, but not near as much as the overall country’s population (and therefore local population density is probably even more correlated).

Correlation Visualization – Best way to read is to look down (or across) the “slope” column/row. The more yellow colors are high correlation with the severity and the darker blue colors are less (or negatively) correlated. The zero color looks a whole lot like the color of “totalSmokingRate’.
Correlation Results Below
slope 1.000000
pop2020 0.141106
maleSmokingRate 0.069965
totalSmokingRate 0.028203
femaleSmokingRate -0.052708

Leave a Reply

Your email address will not be published. Required fields are marked *