COVID-19 US State Data Update – 4/3/2020

For Data Nerds… All the critical parameters for the top US States by Deaths per 1000 citizens – 4/3/2020

Update for Data Nerds. Here’s a table that you might be able to stare at for quite a while. This has all the key parameters that you might care about and give you the data to mentally refute the News Anchor throwing out bad or misconstrued analysis. I’ve heard quite a bit of this.

My Analysis:

Here’s what I see…

  1. New York has crazy high numbers across the board. It’s not simply a matter that they’re earlier into their outbreak. At least as far as we know, they’re in the same boat with Washington and California time-wise. Their situation just got far worse for some reason. Their deaths per 1000 people is about twice as high as the next highest (Louisiana). I’m guessing Louisiana is in their situation due to the extensive hospitality industry they host in New Orleans. Lots of travelers from all over the world, and, oh, Mardi Gras was right at the start of the US outbreak. But New York is quite an outlier. Still not as bad as Italy or Spain, but compared to the rest of the US States, it is first in Cases per 1000, Deaths per 1000, and the Confirmed Case Rate of Change (IROC) feature. This means, they’re worse and getting even worse faster than any other state!
  2. Louisiana’s Death Rate is getting less steep. This could be a one-day anomaly, but the delta IROC also shows that the slope has stayed consistent for at least 2 days. I’m hoping this means that the acceleration of their death rate has ended.
  3. Vermont’s Death Rate, which was high for a while (small states can see faster impact from seemingly-small numbers of deaths), has at least temporarily slowed. You can see this by the dIROC-deaths going to zero and the IROC-deaths being nearly zero. We’ll keep watching this, but I have noticed a trend of smaller states and countries having really bad-looking statistics for a few days and then the outbreak peters out. San Marino may be another example of this. Washington also seems to be in the same category. They’ve been hit hard early in this outbreak. Hopefully this is a good sign and not a short-term data anomaly.
  4. New Jersey has the fastest accelerating curve of normalized deaths over time. I’ve noticed this a couple of times from New Jersey… they accelerate then slow down. Could be bad data capture/release practice or could be something else.
  5. The number of cases are accelerating for all states. What this means is that instead of something like 20 new cases each day (which would be a constant slope of 20 cases per day), all states appear to be seeing acceleration like 20 new cases, then 30 new cases, then 50 new cases, then 100 new cases. This is probably completely normal due to the way viruses spread. Each viral infection results in 1.2 to 2.5 new infections. Viruses always top out somewhere, though (none infects 100% or even 50% — Spanish Flu came close — of the population). 15% is a number that is common for Flu epidemics. So when we see COVID-19 top out, we should see the accelerations of the Confirmed Cases go to close to zero and the IROC for confirmed cases go close to zero. Keep watching for that.
  6. Deaths seem to be staying low or even slowing across the states. This is also reasonable. Lets keep watching to see if this changes for the good or for the bad.

2 Replies to “COVID-19 US State Data Update – 4/3/2020”

  1. Tod – I enjoyed your analysis. Will you speak to the absence of new cases in China and Iran in the last few weeks? Also, China reported today that they have a small number of new cases which suggests that they are still communicating their data. I have been watching this website closely but I have not been capturing the day to day changes in a spreadsheet. Do you know if these daily changes are being captured and are accessible?

    Take care and be safe.

    1. Hi Jim, China and Iran are not being honest with their data… it’s easy to see this when comparing their cumulative rates with those of any other country. I also am noting people backing off of what was “learned” during the original China outbreak because of lots of statistical inconsistencies.

Leave a Reply

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