Continuing with my multiple bar chart races, I decided to create one comparing COVID deaths in the US to other major casualties in US history. I won’t go through all of the code because it is no different than the other bar charts I made. Instead, I’ll just show my sources and a rough idea of my calculations.
I obtained COVID deaths the same way as for past bar charts, but just filtered the data to US only.
There are of course numerous casualties one could include in this analysis. I decided to stick with the ones most people would know. Time Magazine has a good list that I used.
The casualties it lists are: 1. American Revolutionary War 2. 9/11 3. H1N1 2009 4. Korean War 5. Vietnam War 6. 1968 Flu 7. 1957-58 Flu 8. WWI 9. WWII 10. American Civil War 11. Spanish Flu
I looked up the dates for all of those events. I honestly legitimately was not sure how to handle an event like 9/11 that happened on a single day. The other events were easy. I calculated the number of years over which they occurred. For 9/11 I wasn’t sure whether to spread the deaths over one year, put them all on day one on the bar chart, or actually, because they happened all in one day, multiply them by 366. I decided since no one would be satisfied no matter what I chose, that I’d include all 3. So one is 9/11 total deaths/366. One is 9/11 total deaths all on January 1st. And then one is a 9/11 every day.
For the others, when I calculated years, it was inclusive. So, for example, the Revolutionary War. It went from 1775 to 1783. I calculated that as 9 years. The main reason as that if you don’t do that, other events that only occurred over 1 year would come back as 0 years.
I also decided that it was reasonable to calculate the deaths as an average over the duration of the event. For example, take Vietnam. According to the US Archives, the total casualty count in that war was 58,220. The peak year was 1968, when 16,899 Americans died, a rate of 46 per day. The 58,220 were averaged over the 21 years
For population adjustments I relied on three population data sources. For historical population I used this source. Because it was only reported in 10 year increments, I just ran a trend between the years to get estimates for individual years. I then calculated the average population over the years the event took place. That was in order to convert all the death figures to 2020 population. Otherwise, the older the event, the more disadvantaged the numbers are in competing with COVID. For dates after 1950 I used the Worldometers Population chart for the US. Finally, for the 2020 projected population, I used the US Census Bureau Population Clock.
From there I was able to calculate an average annual US population for each event, which allowed me to calculate a multiplier to bring the death counts up to a 2020 population level. I then divided each event by 366 (2020 is a leap year) and put them into the CSV for the whole year and uploaded that to Flourish.
Of note, for WWI and WWII, I only considered the dates in which the US was involved.
This comparison is not perfect. It is hard to justify the comparison of an infectious disease to combat deaths, or deaths caused by a terrorist attack. It most certainly is not meant to minimize the loss of life caused by those events. Instead it aims only to provide context for the magnitude of death caused by COVID by comparing it to events that many will have a conceptual grasp of in terms of the level of death they caused.