GoogleCloudPlatform / covid-19-open-data

Datasets of daily time-series data related to COVID-19 for over 20,000 distinct locations around the world.
Apache License 2.0
471 stars 130 forks source link

Wrong hospitalization data for UK #232

Closed l2edzl3oy closed 3 years ago

l2edzl3oy commented 4 years ago

Repo has the following hospitalization data (and only for Scotland):

date key new_hospitalized total_hospitalized current_hospitalized new_intensive_care total_intensive_care current_intensive_care new_ventilator total_ventilator current_ventilator
... ... ... ... ... ... ... ... ... ... ... ...
1/9/2020 GB_SCT 264 111212   6 7699          
2/9/2020 GB_SCT 258 111470   5 7704          
3/9/2020 GB_SCT 259 111729   4 7708          
4/9/2020 GB_SCT 258 111987   4 7712          
5/9/2020 GB_SCT 251 112238   4 7716          
6/9/2020 GB_SCT 244 112482   4 7720          
7/9/2020 GB_SCT 256 112738   5 7725          
8/9/2020 GB_SCT 267 113005   6 7731          
9/9/2020 GB_SCT 274 113279   6 7737          
10/9/2020 GB_SCT 266 113545   7 7744          
11/9/2020 GB_SCT 269 113814   8 7752          
12/9/2020 GB_SCT 261 114075   8 7760          
13/9/2020 GB_SCT 259 114334   7 7767          
14/9/2020 GB_SCT 264 114598   7 7774          

It does not align with the official UK data (link), e.g.: image

There is ventilator data from that website as well - wondering if it could be captured.

I'd be happy to contribute (i.e. code) if that works. Thanks!

owahltinez commented 4 years ago

Does your recently merged PR fix this issue, or should we keep this open?

l2edzl3oy commented 4 years ago

Does your recently merged PR fix this issue, or should we keep this open?

My PR was for UK country-level data (GB key) only. It subsumes Scotland (GB_SCT key), but I didn't overwrite any existing Scotland data pipeline code or output data.

The source I used (from UK govt) has different Scotland numbers vs the existing Scotland numbers in this repo however. The discrepancy has not been resolved, so I think this issue should still remain open.

Any suggestions on what should be done? I'd be happy to update the existing Scotland code to use UK govt data if it makes sense.

themonk911 commented 4 years ago

I think it makes sense to use the official source, you could also add the other nations at the same time.

owahltinez commented 4 years ago

Sounds good. We still need to use the existing data source for admin level 2, right? Or does the API provide that data as well?

l2edzl3oy commented 3 years ago

So sry for the delayed reply! Had to attend to other work.

Thanks for the confirmations - I have added the other nations (England, Northern Ireland, and Wales), and replaced the existing Scotland data source to use the UK government data source. See #283 .

@owahltinez Nope no need to use the existing data source for L2 data. The API provides both L1 and L2 data!

l2edzl3oy commented 3 years ago

Believe this issue can now be closed since PR #283 has been merged and closed :)