Is time_series_covid19_confirmed_US.csv data cumulative confirmed or daily changes?

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

Is time_series_covid19_confirmed_US.csv data cumulative confirmed or daily changes? #1823

Open MMogollon11 opened 4 years ago

MMogollon11 commented 4 years ago

When I click on time_series_covid19_confirmed_US.csv, my screen shows a data coma delimiter information, not a table as the Global data. I looked at the first row, Alabama, all zeros except that at one point is 3 and the next day is 1 and the rest is zero. If the data is cumulative Confirmed cases, Alabama should be one day 3, next day 4 (3+1) and the rest of them should be 4.

The first line, Alabama, is not the only place where this is happening.

Unless the data is daily confirmed cases, which is different as it has been all this time, and it is now for Global.

skymoore commented 4 years ago

Would be great to get an answer on this one.

mikehumphrey commented 4 years ago

Looking at the ReadMe located at the root of the csse_covid_19_data folder, I offer the following:

This file contains the 'daily time series summary tables, including confirmed, deaths and recovered.
All values in this file are a summary of the values in the daily case report files.
Based on my understanding, these values could be described as the Daily Authorized Value for that location for that specific day. This value is not truly accumulative as the sum of all 'new' cases will not always match the daily authorized value (reference needed). I can not speak to the methodology or process of authorizing these values, but they appear to be the historically correct best value based on the available data sources for the region.

ouinouin commented 4 years ago

Hi, this issue is very big, while trying to do some synthesis with this datasets, you ll see that at least the global datasets mixes cumulative (italy for example) values with daily values (france for example) , this makes it impossible to automate some stats. https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv is it a caveat from the source of information, or the way you collect it ? regards.