The TSA charts are update here.
NOTE: Projections are hypothetical and speculative.
Projections with outlook beyond 30 days are extremely hypothetical.
Peruse cautiously, and use responsibly.
Disclaimer: "I'm with the tour group."
I'm reluctant to add any commentary to the traffic data as the numbers and charts are self explanatory. The last update was June 22,2020. Since then, as the circus maximus of the "re-opening of the economy" across the nation began, we saw a brief surge in traffic that nicely bent the polynomial (with an R2 of 94% to go with it) - but it did not last. It has completely faded.
Traffic was on a nice surge, it hit a high over the July 4th holiday, stalled, and has been moving sideways since then.
The seemingly important crossing of 1 million daily passenger traffic (of which UAL approximately gets 17%, per one of the town-halls) has been delayed:
- the early June projection indicated end of July or beginning of August.
- the current projection now looks like September, 2020 (if traffic starts climbing).
Of course, if traffic does not start climbing, then the September projection will be meaningless.
Since Excel projections looked a bit suspect, I started looking at the comparison of actual traffic and the previous Excel prediction from mid June. You can see the solid sideways movement, below the projection - brown bars (actual) vs orange bars (projection).
Since that comparison graph is running out of projection days, here is another one with a 30 day outlook. Excel uses all of the data, since April 4, 2020, to make that climbing projection. It's dubious - the actual data may well continue to drive sideways for some time to come.
The poly charts are explained in detail in this blog post.
Let's start with a polynomial from the last update (6/22/2020).
Note: I no longer call them "Forecast." The regression (or trendline) is a description of current data.
Compare that nice curve above to the latest polynomial regression below, also 60 day outlook. The new poly now looks like a linear regression, and the R2 is down to 0.9169 (from a high of 0.94), as you would expect after the previous surge pattern changed.
With the traffic pattern no longer being homogeneous, the polynomial equation describing the daily traffic now has a tougher time accurately describing the overall pattern that includes the current sideways movement. The "fitment" is reduced to an R2 of 0.92.
The projection for 8/18/2020 is down from 60% to about 35% - a dramatic shift:
The poly with 120 day outlook is equally flat, and does not project well for Q4.
Here is the June 22, 2020, poly:
You can see that the 120 day poly is also down by about 50% for Q4:
The polynomial for the percentage data (the red lines) is equally flat, and sad, and also down by about 50%:
And finally ...
After not budging much for weeks, the slowly changing linear regression now shows a change of about one month, compared to the June projection:
You cannot read too much into all of this, as all these projections of current data are looking outrageously too far ahead and do not include the million variables that can and will impact actual traffic data.
Asked the man for a few prayers.
The before-and-after reading this blog ...
And - perhaps no longer such a dumb question: