Creating small multiples of COVID data to help form my opinion
Using Tiny Charts To Form My Opinion
As I hear news stories about COVID-19 trends, I like to look at data to form my own opinion if the narratives are in line with the data. I often want to compare current numbers to the high-water marks for bellweather countries and states. For international COVID data visualizations, my favorite site is Financial Times’ Coronavirus Chart.
I like to see the metrics per 1M people so I can compare countries, regardless of their size. Here are the countries I usually compare:
Italy as the one the earliest and hardest hit
Sweden as a country that took a different approach
Canada as our neighbor to the north
New Zealand as a country that took aggressive measures
For state data, I really like The Atlantic’s COVID Tracking Project (CTP) visualizations. They use a nice national trend line when looking at the data per 1M people.
Even with these nice sites, I kept craving a state-based, small multiples chart to allow me to compare states to each other along with the narrative of the virus in the US. Small multiples are simple graphs using the same X and Y axis maximum values to allow you to scan across them to find outliers and make comparisons. They avoid typical chart must-haves like axis labels and tick marks allowing you to focus on relative comparisons. I wanted a chart to look theses states:
NC as my home state
NY as the first and hardest hit
CA as the biggest state
AZ and FL as early opening states with significant surges
DC, MD, and VA as the area where I group up
I finally caved and wrote a script to create variations of small multiples of the CTP data. I colored the lines for states trending up and down in the last week to help with finding current trends. It’s led me to some interesting observations:
The Dakotas and Montana are worse now in new positive cases (per capita) than NY, AZ, and FL ever were.
DC had a NY-like trend in new positive cases (per capita). That was not a big story nationally (at least not in NC) but that might be due to the news cycle focus on racial injustice during that time.
AR has a surprising high amount of recent deaths per 1M people.
Hospitalization data is hard. Many states didn’t start collecting this data until later in the pandemic leading to late and odd data spikes.
Not surprisingly, the most total new cases (not per 1M people) are in the states with the biggest populations: CA, TX, FL, and NY.
If you find the charts to be informative, you can view them on my website and see the code on GitHub.