CSSEGISandData / COVID-19

Novel Coronavirus (COVID-19) Cases, provided by JHU CSSE
https://systems.jhu.edu/research/public-health/ncov/
29.14k stars 18.46k forks source link

Turkey data artifact #3454

Open GrayAnderson opened 3 years ago

GrayAnderson commented 3 years ago

So, Turkey reported over 800k cases in the last 24 hours. Unlike most other artifacts, this one not only seriously distorts the daily chart for Turkey but also makes a dent in the global chart: Turkey: image Global: image

I don't know what the right answer is, but could we please do something to "squeeze out" this artifact from the graphs? Perhaps manually capping the Turkey graph at 50-80k and the global graph at 800k or 1m per day? The impact is significant for the global graph but it makes the Turkey graph utterly unreadable.

TVTvlo commented 3 years ago

チョオン

Vào 23:13, Th 7, 12 thg 12, 2020 GrayAnderson notifications@github.com đã viết:

So, Turkey reported over 800k cases in the last 24 hours. Unlike most other artifacts, this one not only seriously distorts the daily chart for Turkey but also makes a dent in the global chart: Turkey: [image: image] https://user-images.githubusercontent.com/65295193/101988825-3b0f7200-3c6a-11eb-8403-929a37bb0941.png Global: [image: image] https://user-images.githubusercontent.com/65295193/101988849-61351200-3c6a-11eb-9124-87f1e9fa4608.png

I don't know what the right answer is, but could we please do something to "squeeze out" this artifact from the graphs? Perhaps manually capping the Turkey graph at 50-80k and the global graph at 800k or 1m per day? The impact is significant for the global graph but it makes the Turkey graph utterly unreadable.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/CSSEGISandData/COVID-19/issues/3454, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARZXSJDI7PPWPVI6TTYY7CLSUOJEDANCNFSM4UYKK6NA .

GrayAnderson commented 3 years ago

My apologies, but I don't speak Japanese.

flattedk commented 3 years ago

It seems to be an issue that arose in the source data - it's even worse on the Turkey national information page - they have 1.2M cases added on Thursday https://covid19.saglik.gov.tr/EN-69532/general-coronavirus-table.html Worldometer does not show that artifact but still agrees with the Fri/Sat totals