### 1. COVID19 data base

I have been tracking U.S., State, County data in an effort to see if we are making progress reducing the growth rate of confirmed cases, etc.

My projected data uses the latest 7 day moving average of the Daily Growth Rate.

Any Suggestions?

```Current Table: US
#20: {2020,3,31,23,0,0}
CONFIRMED: 189,510 (+25,236) = 1.153 Daily Growth Rate
Daily Grow Rate moving average, period = 7 days: 1.1938
DECEASED: 4,076 (+1,036) = 0.02150 X Confirmed Cases
RECOVERED: 7109
ACTIVE: 178325

Current Table: US
Daily Growth Rate of confirmed case
average over the last 7 days: 1.193  -- 1.193886987
Ratio Confirmed cases/deaths: 0.02151 = 2.151%

Daily Growth Rate Moving Average, Period = 7 days:
3.095
3.105
3.128
1.209
1.391
1.365
1.334
1.294
1.271
1.249
1.234
1.199
1.194  -- 1.193886987

based on a Daily Growth Rate of: 1.193
display("based on a Daily Growth Rate of: [.3g]", mva[\$])

based on a Daily Growth Rate of: 1.194
printf(f,"based on a Daily Growth Rate of: %1.3f\n", mva[\$])
1.193886987

Coronavirus: US Projected Data
Day     Date        Confirmed       Deaths
#1      2020-4-1:   226,254         4,866           +790
#2      2020-4-2:   270,122         5,810           +944
#3      2020-4-3:   322,495         6,936           +1,126
#4      2020-4-4:   385,023         8,281           +1,345
#5      2020-4-5:   459,674         9,887           +1,606
#6      2020-4-6:   548,799         11,804          +1,917
#7      2020-4-7:   655,204         14,092          +2,288

```

### 2. Re: COVID19 data base

Hi fits with a doubling every 4 days (1.19 ^ 4)

Are you trying to model it?

### 3. Re: COVID19 data base

ChrisB said...

Hi fits with a doubling every 4 days (1.19 ^ 4)

Are you trying to model it?

I suppose I am trying to model it. I am out of my league, as usual.

My notion was that I might be able to determine a +- growth factor somehow from the moving average of growth factors. which I could use to progressively (recursively?) project future data points. But I am stabbing in the dark. Just trying stuff.

Regards, Ken Rhodes

### 4. Re: COVID19 data base

Yeah, huge number of factors covering the infection rate of this virus, start by listing the factors that cover it's spread, and you can start to see the epidemiology rabbit hole you're falling into.

Good luck

Cheers

Chris

### 5. Re: COVID19 data base

To see the growth you will have to remove cured patients, to keep only infected patients.

Then you can plot a graph to see the different phases of the pandemic.

Jean-Marc

### 6. Re: COVID19 data base

jmduro said...

To see the growth you will have to remove cured patients, to keep only infected patients.

How do you account for the asymptomatic infected non-patients who never get tested, but keep spreading the virus? Or those with mild symptoms, who chalk it up to normal seasonal allergies? Or the varies-by-state practices of "report all deaths as covid, and disprove it later" vs "don't report any deaths as covid unless certified by testing and lack of co-morbidities"?

EDIT: Also, there's still a large backlog of tests (1000's in a few places i bothered to look), and some of those tests may be for people who have already died. This is leading to people being counted as dead and days later being counted as newly infected. Etc..

A browser-based gui for the John Hopkins University covid database. Nope, i did not say it is the best, the most appropriate, written in OE, or anything else, just that it exists.

Curious as a...
Kat

### 7. Re: COVID19 data base

Thanks for all the comments.

I obtain my U.S.A. data from: WHO and data the GA (USA) data from: GA Georgia does not report "Recovered" cases.

I also refer to: Health Data which makes projections based upon various assumptions:

```We classified social distancing measures using the New Zealand Government
alert system Level 4 and then assume that locations that have instituted
fewer than three of these measures will enact the remaining measures within
seven days. We also assume that implementation and adherence to these measures
is complete. With each model update, the assumption of full implementation of
social distancing measures is reset; any delay will be reflected in the number
of deaths and burden on hospital systems that the model estimates.
```

Based upon the assumptions cited above, I think the healthdata.org projections/forecasts are significantly biased toward the optimistic.

I am going to try to avoid the "epidemiological rabbit hole" by keeping it very simple while realizing that many caveats apply:

```Current Table: GA
Day 22: 2020-4-2
CONFIRMED: 5,444 (+696) = 1.146 Daily Growth Rate
Daily Grow Rate moving average, period = 7 days: 1.1907
DECEASED: 176 (+22) = 0.03232 X Confirmed Cases
RECOVERED: 0
ACTIVE: 5268

The Daily Growth Rate Moving Average for Confirmed Cases
has declined for 9 day(s).
Ratio Confirmed cases/deaths: 0.03233 = 3.233%
```

I believe the moving average gives a pretty good approximation as to whether we are moving in a positive or negative direction.

Thanks again for all the comments and suggestions and any more will be welcomed.

Regards,
Kenneth Rhodes

### 8. Re: COVID19 data base

```Current Table: US
Day 23: 2020-4-3
CONFIRMED: 277,953 (+32,740) = 1.133 Daily Growth Rate
Daily Grow Rate moving average, period = 7 days: 1.1509
DECEASED: 7,152 (+1,169) = 2.573% of Confirmed Cases
RECOVERED: 9,823
ACTIVE: 260,978

Moving Average Period = 7 days

Confirmed Cases:
The Daily Growth Rate Moving Average has  DECLINED
for 11 consecutive days

Deaths/Confirmed Cases:
Deaths/Confirmed Cases Moving Average has  INCREASED
for 9 consecutive days

```

### 9. Re: COVID19 data base

Some of this information is already available. For example, https://ourworldindata.org (click on the narrow orange banner near the top, or scroll down to find the COVID-19 section) has a lot of data in table, map and chart form, from sources that appear to be reliable. Well, as reliable as any other, at least.

The interactive charts in particular are well designed - data can be selected by date, country, region or a combination.

### 10. Re: COVID19 data base

Way over my head...

UPDATE: added a simple solution: https://rosettacode.org/wiki/Logistic_Curve_Fitting_in_Epidemiology#Phix