Day 30: No Big Tuesday Bump, but Not Great

• COVID-19 Report As of Tuesday, April 14, 2020 GA DPH Noon
• Confirmed COVID-19 Cases: 14,233 (+733) ▼
  • Consecutive day #1 of decreasing growth (14 needed) ▼
  • Daily Growth Rate: 6.8% (-0.4%) ▼
  • Double Number of cases every 15 days ▼
  • Average Growth Rate (7-Days): 7.0% (-2.1%) ▼
  • % of Population Affected: 0.1358% (+0.0087) ▲
• Estimated
  • Total Infected (Est.): 142,230 (+9,080) ▲
  • Contagious, No Symptoms (Est.): 111,105 (+7,174) ▲
  • Contagious Ratio (Est.): 11 in every 1,000 Georgians (+1) ▲
• Hospitalized: 2,769 (+180) ▲
  • Beds Available: 8,322 ✓
• Deceased 501 (+37) ▲
  • Morbidity Rate: 3.51% (+0.03%) ▲
• ▼= getting better; ▲= getting worse; ✓ = same

The best news is the average rate of increase continues to decline to 7.0%. Fig 1, above, orange line. As that rate continues to decline, we can hope that the virus is burning itself out. However, the number of infected Georgians continues to rise. Fig 2, below. The numbers show continued growth, but slowing growth. The predictions are that the numbers will continue to rise (that is the very definition of growth after all). But this continued decline in the average rate of growth tells me that the shelter-in-place / social distancing is working because everything else has remained the same (same disease, same contagion rate, same population size). But those 11 out of every 1,000 Georgians who are possibly running around infecting others can be the undoing of it all.

CAUTION MODEL GEEK INFORMATION AHEAD: I intended to leave my model predictions alone, but apparently Excel had other ideas. Somehow I lost the entire graph and had to restore the file from yesterday’s backup. When I did that, I noticed that I had some errors that I just had to fix, and before long, I was in for a penny, in for a pound. The attentive reader will note that the prediction lines are different from those presented yesterday. I was clearly not trying to rewrite history to make my predictions seem better because I made them appear worse. However, going forward I will (do my best to) leave the old predictions in place and will adjust the predictions daily–sort of a combination of what I started out doing. My rudimentary models consists of an average growth for the whole epidemic (yellow line) that will be adjusted daily as the average changes (currently 407 per day but previously lower). It looks straight today, but it will get some kinks in it as the average changes. Next is an exponential growth curve that should be the most accurate using the average growth rate for the last 7 days (orange line). This is a “rolling” average that will also change from day-to-day. The green line is an average of the yellow and orange lines with an added an x-factor. This x-factor is the average of the errors of the other two models from the preceding day, so the average can be positive (overestimation) or negative (underestimation). I expect the average to be low and the exponential to be high and the green in between. These are the same models that I’ve been using, but last week, I did not allow adjustments (or I did not intent to allow adjustments, but copy-and-paste did). (I should probably use a weekly average error instead of the daily average error, but I’m glad to just get to where I am now considering the mess I had to fix to get here. It was working so I did not want to mess with it too much. Now that I’ve messed with it, I might go weekly average tomorrow. Who knows.)