Read on Twitter
STOP mapping COVID case counts by zip code Zip codes come in varying shapes and sizes, introducing bias known as the Modifiable Areal Unit Problem (MAUP).
These two maps both display the # of recorded COVID cases in NYC. Very different takeaways.
#gischat #COVID19
The map on the left is a simple choropleth map displaying raw counts of recorded COVID cases by zip code from @nycHealthy
The map on the right is a dasymetric map which redistributes raw counts of recorded COVID cases from zip codes to residential buildings proportionate to building volume, and then re-aggregates by equal area hexagon.
COVID case counts should be accompanied by data on the total number of tests administered. It is flawed to draw conclusions about the spread of COVID without accounting for the number of tests administered. And it is flawed to aggregate this data by zip code. #gischat #COVID19
There is some sloppy research right now which make conclusions based on overlaying data on top of raw COVID case counts by zip code, such as this paper from @jeffEconomist which suggests that subways seeded the COVID epidemic. This is seriously flawed! https://www.nber.org/papers/w27021.pdf
Code (as a Jupter Notebook) available here: https://github.com/willgeary/covid-nyc-dasymetric-map
That's pretty harsh. Perhaps consider phrasing questions and feedback about people's work in a kinder way? The color scheme used here is called "Viridis" and there is ample scientific research pointing to this scheme demonstrating high speed and accuracy: https://idl.cs.washington.edu/files/2018-QuantitativeColor-CHI.pdf
