Saturday, September 25, 2021

High Vaccination, Low Hospitalization

Kenneth Field has posted a nice critique of a Washington Post bivariate choropleth map showing America’s hospitalization and vaccination divide.The Post's bivariate map uses colors to visualize two different variables on one map - the Covid-19 vaccination rate and the Covid-19 hospitalization rate in every U.S. health region.

Kenneth's Cartoblography post praises the map for making it very clear that "Regions with more vaccinations have fewer hospitalizations." This is where I have to make an embarrassing carto-confession - I often find bivariate choropleth maps difficult to read. When I first looked at the Post's map I thought that the dark colors in Texas, Florida and Oregon identified these areas as having high rates of vaccinations and also high rates of hospitalizations. I also read the map as saying that the light colored Kansas has a low rate of vaccinations and a low rate of hospitalization.

The Post's article also includes a scatterplot of the same two variables which I find a lot easier to read. Looking at this scatterplot reveals that Kansas probably is an outlier (although because the plot isn't interactive I am having to guess that one of those two dots in the bottom-left segment indicates Kansas). However this scatterplot also reveals that there are no regions in the highest vaccination / highest hospitalization sector (however it does appear that parts of Oregon and Florida are in fact close to that sector)

I think that I find the Post's bivariate choropleth map hard to read because it has 16 different colors and it isn't just a matter of light hues = good / dark hues = bad (or vice versa). As Kenneth points out in his critique the Post does help the reader by using annotations both on the map legend and on the map itself to help the reader understand the data.However I do find that the use of so many colors on bivariate maps means that I often have to work towards comprehension. That in itself isn't necessarily a bad thing but it might be a factor worth considering if you want to visualize two different variables on one map.

Obviously the Post's accompanying article also makes clear the correlation between vaccinations and hospitalizations. The data is unequivocal - vaccinations work. Areas of the country which have the highest vaccination rates tend to have the lowest hospitalizations for Covid-19. Conversely those regions of America with the lowest vaccination rates tend to have some of the highest hospitalization rates for Covid-19. 

The message couldn't be clearer. If you haven't done so already then you need to get vaccinated! 

No comments: