Happy Data Cleanup Day (aka Saturday). King County reported 122 infections, almost half of which were spread out over the past few weeks. So much for clean data. Also, I’ll continue to note that the county does not provide us common folk data on negative tests and total tests administered, among other missing data points.
Even though the model was off by half of today’s number (predicted 65), because the infections were spread over the past few weeks, the model still looks good, retained its full predictive power (r-squared of 0.7741), and is going with 63 tomorrow.
We’ve seen some Sundays have data cleanup as well, and sometimes we get the straight infection number from the previous day. Not sure what’s going to happen tomorrow. I’m sticking with the model. Look at that curve!