Number of hospitalized patients is a floating point number

[eguidotti]

Hi Matt, hope you are doing well!

@MinerMax555 noticed that for some dates (example: 2021-01-01) the number of hospitalized people in Australia is a floating point number (20.44 on 2021-01-01).

• https://raw.githubusercontent.com/M3IT/COVID-19_Data/master/Data/COVID19_Data_Hub.csv

Do you know what this floating point number means? Is it a data mistake or am I allowed to ruond it to the nearest integer?

Please see also the corresponding issue on the data hub repo:

• https://github.com/covid19datahub/COVID19/issues/193

Many thanks!

[M3IT]

Thanks for bringing this to my attention Emanuele.

I've looked into it and the reasons are:

For hospitalisations, in some states the numbers provided are:

• number being treated (Includes hospital at home)
• % of those who are hospitalised

So the precise number of hospitalisations isn't known, but is calculated from these 2 numbers.

For vaccines, second doses are provided by some states as % of population, so again the number is calculated.

I've fixed this by rounding all counts to the nearest integer and then coercing to integer so they stay that way downstream (and added a check). This fix is in tonight's update.

[eguidotti]

Hi Matt, that's awesome! I confirm it is working fine now. Thanks a lot for taking care