Quick Calculation for Projected New Coronavirus Cases (USA Only)

Nick Austin
2 min readMar 12, 2020

I fitted a natural exponential function to the USA’s daily reported cases of novel coronavirus. Unless drastic measures are taken, we could see 90K+ cases before end of March 2020.

This shows an exponential function fitted to the data.
I linearized the predictions to calculate an R-squared. High R-squared (0.97) suggests model is reasonable.

The model estimates 80K — 90K reported cases in USA by March 24th (assuming we get enough test kits to even administer the tests and report the numbers).

I’m ignoring the leveling-off effect of population growth here to model out a pessimistic scenario. We could attenuate the growth with drastic social distancing measures.

After some time the curve will level off to resemble an S-curve (at which point the growth becomes logarithmic rather than exponential) but we’re very far from that inflection point. I wouldn’t even think about inflection until we’re well into the hundreds of thousands (or millions?).

Let’s hope the model is wrong.

Caveats: This model only looks at reported cases. The number of absolute infections in the USA will be much higher since many infections are asymptomatic, and not everyone who has the virus will be tested.

Data source: https://www.worldometers.info/coronavirus/country/us/