Lockdown Latitude

It was thirty days around the horn
The captain says it's thirty five more
The moon looks mean and the crew ain't stayin'
'There's gonna be some blood' is what they're all sayin'
--Jay Ferguson

Interesting regression analysis here. First graph is estimated change in social mobility over past week for each state (a proxy for relief of lockdown status) vs statewise % positive CV19 tests in past 7 days. No significant relationship between these two variables.

No relationship between #US states with increase in #COVIDー19 % positive tests vs increase in google maps social mobility. Significant relationship with latitude of state capital and total deaths to date. Combination of latitude + total deaths best predictor of new + tests pic.twitter.com/G1rf60Kn3S

— TheShakespeareanApe (@alexkx3) July 5, 2020

Next graph is state death rate (# deaths/state pop) vs statewise % positive tests in past 7 days. Slight negative relationship which would be much higher if asymptotic curve was fit to the data rather than straight line. What this is saying is that states with low total death rates are those with the highest recent positive tests.

Final graph is geographic latitude of state capitol vs % positive tests in past 7 days. Significant negative relationship with R2=.31. The farther south the state, the greater the recent increase in case count.

Last panel summarizes estimated statistical relationships from various combinations of independent variables on the % positive in past 7 days dependent variable. Increase in social mobility explains little or no variance in case count increases. The best combination is state death rate + state capitol latitude. Significant negative relationship. R2=.39. p-value very low.

A sampling of some of the outstanding analysis that has been surfacing on social media. Also disturbing that public health agencies and mainstream media demonstrate little aptitude for such quality work.