CSSEGISandData / COVID-19

Novel Coronavirus (COVID-19) Cases, provided by JHU CSSE
https://systems.jhu.edu/research/public-health/ncov/
29.12k stars 18.4k forks source link

Structural fix on major Countries Recoveries stats (e.g. US, France) is BADLY needed #1230

Open r-lomba opened 4 years ago

r-lomba commented 4 years ago

I always start my posts with my appreciation for the hard work at JHU in maintaining this valuable stream of data, but:

Those are major Countries, and it's ok to have some holes in the stream of stats for a few single days, but here we are talking about a source of data used by thousands of people to assess a pandemic evolution and, if it stays like this, it's unreliable

Major Countries can't have a structurally wrong stream of stats. This must be fixed somehow, and very soon

yetzt commented 4 years ago

For most countries, recovery data is not officially collected, there is no notification requirement. You can only treat this data point as a minimum number.

r-lomba commented 4 years ago

@yetzt you don't get the point

Here, NO data is officially collected. They state that clearly. But this is a service JHU offers to the world, it's not that I'm a nerd pointing out a wrong number in a cell.

I'm pointing out that this is an unreliable service as long as the stats related to major countries are structurally wrong, and people gets frustrated when a service is unreliable, and also you as an institution are providing a BAD service to the community

For instance, how comes that these guys here do have the stats for US and France Recoveries?

https://www.worldometers.info/coronavirus/

So, notification is not only required but mandatory (if one wants to be constructive) when there are major flaws like these I've pointed out. It's of course JHU that will evaluate how important this is with regards to their policies and goals and expectations from this project

JiPiBi commented 4 years ago

Worldometers displays correct values on recovered for France : they use official sources or media sources and give the link to the document or video they used , quite good job

JiPiBi commented 4 years ago

But worldodometers is neither perfect , they sum up twice the territories of France (1st in the line France which is a total agregation , and 2nd in the territory itself) I made a second message to them , because they didnt change after the first and neither replied to me

yetzt commented 4 years ago

@r-lomba i'm not with JHU, i'm a data nerd pointing out wrong numbers in cells, but also trying to explain things. ;)

jcavfen9 commented 4 years ago
  • France Recoveries are reported as 12 since ever

I'm not sure French administration is publishing this data in an structured manner (I did not find it). I don't know where Woldometers got the figure. I imagine it's not easy for JHU to retrieve it. Some people are trying to structure French data here.