search

datasets / covid-19

Novel Coronavirus 2019 time series data on cases
https://datahub.io/core/covid-19
1.16k stars 604 forks source link

France aggregated count is down from yesterday, why? #63

Closed efaysal closed 4 years ago

efaysal commented 4 years ago

France count or confirmed number has an issue, 82 2020-04-12 133670 83 2020-04-13 137875 84 2020-04-14 131361 why the number is going down from yesterday? As it is an aggregated number is has to growth or to show stagnation ... Thank for any clarification.

anuveyatsu commented 4 years ago

@efaysal yes, these are cumulative figures and must not decrease. This must be a mistake in the upstream and hopefully will be corrected anytime soon.

efaysal commented 4 years ago

It is very surprising that the processing is not making a check on this kind of issue. A simple check in the code as if last is greater than the next, exactly if diff(x) shows a negative value then stop to fix!

I guess the problem may be related to France territories, main land is in Europe and the territories are scattered in pacific and America and so on. Counting over all territories, something is not making sense ...

However, the counts related to other countries make sense to me: Some numerical results are in https://www.linkedin.com/posts/faysal-el-khettabi-ph-d-4847415_covid19-activity-6656536209562619904-IFVH

anuveyatsu commented 4 years ago

@efaysal this project doesn't do any data modification but standardization/normalization if possible. We also can add more metadata etc.

anuveyatsu commented 4 years ago

Btw, France data was corrected recently - https://github.com/datasets/covid-19/commit/bed2cbaeccabece8ae86e15363cfc910afdbddb2

anuveyatsu commented 4 years ago

FIXED:

2020-04-13,France,125394,28001,14986
2020-04-14,France,131361,29098,15748
2020-04-15,France,134582,31470,17188