Drumroll please… The linear model predicted 161 new COVID-19 infections today… and the result from King County Public Health: 175. Not bad!
A little concerning: of the new tests, 17% were positive. That’s up from 13% the day before. The number of test hit almost a thousand again – most tests conducted in a week.
Once again, I’m very disappointed with the data quality coming out of King County. 14 days of the past month were updated with revised numbers, going back all the way to March 9th. Why did infections from March 9th increase by 3 (almost a 10% increase)? Why did they go up by 4 on March 10th (over 10% increase)? Did the positive tests get locked in a filing cabinet?
Here are the differences reported by day:
The new model is extremely simple and very predictive: It says the number of cases will increase in King County by 5 per day. R-squared value of 0.8132. Not bad
Prediction for 4/2: 166
The model is humming!