Just checking my work here. And it’s a nice illustration that sometimes, “stuff happens” in the observational data. But the conclusion remains the same: There was no post-Super-Bowl explosion of new COVID-19 cases.
Recall Post #1029.
Obvious conclusion: No post-Super-Bowl explosion of new cases. Above, Blue is Tamp, Orange is KC, Gray is the rest of the country.
Case closed? Let’s give it a little more time.
Well, hold on a minute. That’s a little odd. Now it looks like there is a distinct bump-up in KC, just about the right time. (It looks sharper than the graph above because the horizontal scale compresses as I add more days.)
But nothing in Tampa? Florida as a whole is not known for its quiet enjoyment of sporting events, let alone national championship victories.
That nice, symmetrical, week-long increase in the KC data looks suspiciously like one of those many data reporting “speed bumps”. Those are week-long bump-ups in the curve that occur when a state dumps a lot of old cases into the file, all at once. (They last a week because I use a seven-day moving average to smooth the data.)
Because the KC MSA combines KC, KS and KC, MO, let me split by state. The state health department is the reporting entity here. If this is a reporting issue, that should sort it out.
And well-la. Either you believe there was intense partying on the KS side of the line, but quiet sobriety in MO, or you realize that the bump on the KC, KS line is almost certainly a data-reporting “speed bump” that just happens to fall at the right time.
And the state data appear to bear that out. The entire state of KS saw a big increase in the seven-day moving average for that week. Arguably, if you could go back in time, you might even catch the explicit notice for that on the KS COVID-19 dashboard. But those notices are ephemeral, and chances are, if KS explained the increase, that explanation is now lost.
Upshot: The conclusion remains the same. There was no post-Super-Bowl explosion of cases. At first glance, it looks as if there might have been one in Kansas City, but not in Tampa. But splitting that by state reveals that the well-defined, symmetric, week-long increase in cases there is almost certainly just a data reporting “speed bump” from the state of Kansas.