COVID-19 Interest Group
April 6, 2020
Arlin‘s Corner # 2 COVID-19 Interest Group April 6, 2020
The data, information and estimation of future cases presented below is based on public data and my own personal mathematical analysis of it. It has not been peer reviewed and has not been officially endorsed by the Berkeley Rotary Club. It is a brief summary concentrating on today's Covid19 situation in the Greater Bay Area and Alameda & Contra Costa Counties. If you have any questions, please call me at 510-501-0298. If you are interested in a copy of my Excel spreadsheet for the analysis of GBA or A&CC data, please e-mail me at [email protected]......Arlin Peters
===============================================================================================
This was sent out as an email to the BRC Covid-19 Interest Group on April 6th after the 4PM group meeting:
I think Grier said it best during the zoom meeting today, when he pointed out that what I’m calling “estimates”, are in reality my “best guess-timates”. Yes, it’s absolutely true,in face of what limited poor data we have, and all the things we know that we don’t know, and the assumptions that I have to make, that my " best guess-timates" represent order of magnitude thinking. But, in the land of the blind, the one-eyed may become king. And may God protect us from the things of which we have no idea that we don’t know.
I have worked out a new method for “estimating” future cases for the Greater Bay Area (GBA). Instead of using a constant doubling time (DT) for the next two weeks in my “estimation”, I now use an increasing DT over that period based on past GBA data. Just to save typing, I’m not going to put quote marks around “estimate” words in what follows, but they are there nonetheless.
What I did in the past, was to take the data on the number of cases and estimate what the most recent 4 day rolling Doubling Time (DT) was. Then I would use that DT (as a constant) and yesterday's cumulative number of confirmed cases to estimate the number of cases in the next two weeks. For example, the number of GBA cases on April 2 was 2,886 and the most recent DT estimate was 7. I used these numbers to estimate that the number of cases two weeks out would be about 11,500.
This method assumed that the current DT of 7 would stay constant for the next two weeks. This is not correct, DT's are increasing! Analysis of Covid19 data from the Italy, Spain, the US, NYC and the GBA now clearly shows that DT's are increasing and increasing in a predictable linear fashion. What this means is everybody's collective efforts at hygiene and social distancing is reducing the rate of infections, and this is working well over time.
As of yesterday, in Italy, the DT has increased from 5 to 18.5 at a rate of 0.9 DT per day; In Spain, the DT has increased from 4 to 13 at a rate of 1.0 DT per day; In the whole US, DT has increased from 2 to 6.5 at a rate of 0.28 DT/day; In NYC, DT has increased from 4.8 to 7.8 at a rate of 0.35 DT/day; In the GBA, DT has increased from 4.5 to 9.1 at a rate of 0.41 DT/day.
Yesterday April 5, the number of confirmed Covid-19 cases in the GBA was 3623, and the most recent DT was about 8.5. I have used these plus a 0.4 daily increase in DT, to estimate that the number of GBA confirmed cases in two weeks April 19 will be about 8800. In addition, I now can estimate, that the percentage of GBA population who has been infected by Covid-19 yesterday was 0.8% and the % on April 19 will be about 1.7%.
One note of clarification: I am using a multiplying factor of 10 to go from the number of confirmed tested cases (severe cases tested and reported about 7 days from their day of infection) to the total number of cases (including asymptomatic, mild and moderate-cases who have not been tested - plus the severe cases). Based on Chinese data this was a factor of 7. This is perhaps the biggest source of error in my “estimation” and hopefully is on the high side. I welcome any better information or better guess at this factor! So the range of cases for April 19, might best be thought of as being between 0.17% to 1.7%.
I have attached my GBA spreadsheet. Data analysis is on the "Data Confirmed Cases" sheet. In the bottom right hand side is the graph of DT and showing the slope of 0.4. Estimation of future GBA cases is on the "Projected # of GBA Cases" sheet. In the top right is the graph of the estimated GBA population % infected over time.
Stay Safe and Calm - Keep up the shelter in place behavior...................Arlin