Modeling and Predicting the U.S. Aviation Recovery

50% recovery predicted by September; 90% recovery by next March

Over a span of 45 days from March 1st through April 14th, air passenger travel in the U.S. fell by a staggering 96%. The COVID-19 pandemic caused an unprecedented reduction in passenger flights, and turmoil throughout the industry. When will things get back to normal?

The Transportation Security Administration (TSA) publishes data daily showing the number of checkpoint screenings in the U.S., providing a helpful picture of how many passengers are taking to the skies each day. While travel numbers today remain low compared to normal (i.e. 2019 data), there are strong signs that show the recovery is already well underway. Five weeks since that low on April 14 through, daily checkpoint numbers have nearly tripled.

While it’s difficult to predict what the situation will be like many months from now, we can make a mathematic model of the recovery using regression techniques, and extrapolate the model to make a forecast. The model I’ve selected is a logistics function, which is useful for situations involving growth between two levels over time, with a slope of zero at the beginning and end. The recovery is nearly exponential at its start, tapers to linear, and asymptotes to the final target.