Big issues with Poland

[eguidotti]

Hi there,
the data for Poland seem to be cumulative counts before 2020-11-23 and daily counts afterwards (cases, deaths, tests).
The nuts_2 column is also mixed with the same regions referenced with different strings due to encoding issues.
It would be great to fix that by standardizing this file.
That would be really helpful as this is the only source for Poland at regional level I could find!
Many thanks,
Emanuele

[emptymalei]

Indeed. This is confirmed. We should have the cumulative values.

We will need to fix all files in this folder:
https://github.com/covid19-eu-zh/covid19-eu-data/tree/master/dataset/daily/pl

I believe we will need to fix the download and aggregation script too.

Added to my todo list.

[emptymalei]

This should be fixed now.

[eguidotti]

Mmm... it seems that tests are not cumulative and the issue with encodings is still there.
I would like to ask:

• are tests daily tests or should they be cumulative numbers as well?
• the column nuts_2 does not seem to be corresponding to the NUTS codes. Indeed, for Poland we have 17 NUTS2 codes but it seems to me that "Warszawski stołeczny" and "Mazowiecki regionalny" are provided together as the NUTS 1 "MAZOWIECKIE".
• it is difficult to track the same region in time because it is referenced with different names due to encoding issues. Would it be possible to fix that?
• in the nuts_2 column there are wrong entries such as (a) website addresses and (b) some names that do not correspond to any Polish region

Many thanks for your efforts!

[emptymalei]

@eguidotti This is very hard to fix and manage. I decided to stop collecting data for Poland.
9f1c347

[eguidotti]

Completely understandable. Thanks anyway

[emptymalei]

@eguidotti ❤️

If you need the original data, you can download all the data from here

https://arcgis.com/sharing/rest/content/items/a8c562ead9c54e13a135b02e0d875ffb/data

These are the daily data files.

There are some data before this, but as you have already spotted, there are many inconsistencies between the datasets.