Italian Output #6
Comments
|
The issue is that case is inferred from deaths and rounded case fatality.
We need to find a way to reconcile with the Bollettino data, tbd.
…On Mon, Apr 20, 2020 at 11:50 AM Marius D. Pascariu < ***@***.***> wrote:
Checking the Italian output i can see:
Country Region Code Date Sex Age AgeInt Cases Deaths1 Italy All ITinfo15.04.2020 2020-04-15 b 0 10 1008.8 1.02 Italy All ITinfo15.04.2020 2020-04-15 b 10 10 0.0 0.03 Italy All ITinfo15.04.2020 2020-04-15 b 20 10 7061.9 7.04 Italy All ITinfo15.04.2020 2020-04-15 b 30 10 13114.9 39.05 Italy All ITinfo15.04.2020 2020-04-15 b 40 10 19392.1 173.06 Italy All ITinfo15.04.2020 2020-04-15 b 50 10 30103.7 746.07 Italy All ITinfo15.04.2020 2020-04-15 b 60 10 23808.5 2242.18 Italy All ITinfo15.04.2020 2020-04-15 b 70 10 25532.0 6074.39 Italy All ITinfo15.04.2020 2020-04-15 b 80 10 26097.4 7890.410 Italy All ITinfo15.04.2020 2020-04-15 b 90 10 6907.3 1359.011 Italy All ITinfo15.04.2020 2020-04-15 b 100 10 2440.5 976.1
When Age == 10 the number of cases is 0 (unlikely!). I think the cases
here are estimated based on the death records, information which is taken
from daily summarized info-graphics
<https://www.epicentro.iss.it/en/coronavirus/bollettino/Infografica_15aprile%20ENG.pdf>
.
However, note that on a weekly basis complete information by age and sex
is provided (see April 16
<https://www.epicentro.iss.it/coronavirus/bollettino/Bollettino-sorveglianza-integrata-COVID-19_16-aprile-2020.pdf>).
The modeled daily data should be consistent with the complete weekly data
at the nearest calendar date ...or removed.
Also if data is already reported in the desired age classes, I don't think
any altering of the data should be applied. This is the case for the number
of deaths displayed above. Only the deaths in the last reported age group
90+ should be split between 90 and 100. For all the other ages should be
read as reported. I know this complicates a bit the coding but would be the
right way to do it.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#6>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAG43G5SHA2JCWSOZTFVIQ3RNQLGFANCNFSM4MMJL3DQ>
.
|
|
Hello，why is there a decimal in the death toll? |
|
There are 3 things (that I can think of right now) that can lead to
decimals in estimated/harmonized counts.
1. counts of unknown age are redistributed proportionally to counts of
known age
2. if marginal totals are captured, then age-specific counts are rescaled
to sum properly
3. when ages are split from 10 to 5-year age groups a statistical model is
used
We round output to one decimal place rather than to integers in part
because it makes users notice this point. This is in keeping with practice
at other standardized databases, such as the Human Mortality Database.
…On Mon, Sep 28, 2020 at 11:14 AM Nicefyy ***@***.***> wrote:
Hello，why is there a decimal in the death toll?
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#6 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAG43G7KPROCYPHNPM6I4RDSIBHWTANCNFSM4MMJL3DQ>
.
|
timriffe
pushed a commit
that referenced
this issue
Oct 27, 2020
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Checking the Italian output i can see:
When
Age == 10the number of cases is 0 (unlikely!). I think thecaseshere are estimated based on the death records, information which is taken from daily summarized info-graphics.However, note that on a weekly basis complete information by age and sex is provided (see April 16). The modeled daily data should be consistent with the complete weekly data at the nearest calendar date ...or removed.
Also if data is already reported in the desired age classes, I don't think any altering of the data should be applied. This is the case for the number of deaths displayed above. Only the deaths in the last reported age group
90+should be split between90and100. For all the other ages should be read as reported. I know this complicates a bit the coding but would be the right way to do it.The text was updated successfully, but these errors were encountered: