Open GerardoRTobar opened 4 years ago
They have answered this multiple times (see issue #1250 ). The problem is there are no accurate estimates of recovery data it is a BS stat. The bigger issue is the backward-incompatible changes in cases/deaths, which are hard numbers (see my issue #1348 )
@CSSEGISandData the data was already messed up, but consistently messed up, now adding columns with a different level of granularity!?.
Revert and run the old script to process the data using the old rules (I bet on an R script no one can understand but you), if you want to add a new format create another feed that you will have to QA but continue with the old.
This causes less disruption to everyone who is already using the feed and lets you incrementally do whatever you want to do.
On the new structure think about NOT mixing dimensions (cities and counties and cruise ships), and NOT adding columns that represent the same thing for different time periods.
Makes sense? Come on ... you can do it!
For those looking for it. I've format shifted the new data back into the old format. I will be doing this for all subsequently released daily reports.
Check it out here:
Here's a quick and dirty iPython to fix it, needs latitude but Im not using it currently. https://github.com/joaolcorreia/COVID-19/blob/master/iPython/COVID19-JohnHopkins-Fixer.ipynb
Here's my viz: https://public.tableau.com/profile/joaocorreia#!/vizhome/COVID19_15841706004780/COVID-19?publish=yes
I think we all did our kind of ETL process with the data provided in the format before. Besides, the Recovered people information give powerful and needed insight.
Why did you guys top bringing information about the Recovered people? We are not stealing your job, I think we all make proper reference to this site and to your work when using the data.
I kindly ask you to reconsider bringing back the old format. It would be so apreaciated by us and, who knows, by the world
Thanks for comprehending our needs-