search

owid / covid-19-data

Data on COVID-19 (coronavirus) cases, deaths, hospitalizations, tests • All countries • Updated daily by Our World in Data
https://ourworldindata.org/coronavirus
5.66k stars 3.64k forks source link

Negative cases and deaths #66

Closed NigelCusc closed 4 years ago

NigelCusc commented 4 years ago

I noticed that some of the data consists of negative values for cases or deaths. For example GBR United Kingdom 2020-05-21

How is this interpreted?

Thanks!

edomt commented 4 years ago

Hi @nigelc00l

Negative values in cases and deaths can sometimes appear when a country sends a correction to the ECDC, because it has overestimated the number of cases/deaths. For the UK, those –525 cases on 21st May are therefore people that had been misclassified (or maybe counted twice) as COVID-19 cases.

Edouard

HebeHH commented 4 years ago

Is this also the case for Spain's new deaths data for May 25? It just looks very weird.

image

edomt commented 4 years ago

Hi @HebeHH Yes, that's also the same issue (see https://github.com/owid/covid-19-data/issues/84). Unfortunately we don't have enough information (such as the exact days on which too many deaths were counted) to correct past values, so our only solution is to keep the ECDC's –1918 value for that day. Edouard

ospinag commented 4 years ago

great, thanks!

George6328 commented 4 years ago

Why is it the 'total cases and total deaths' do not change with respect to collection date, but the 'total tests' do change?

edomt commented 4 years ago

@George6328 This is because the source and collection methods are completely different for cases/deaths and tests.

Confirmed cases and deaths: our data comes from the European Centre for Disease Prevention and Control (ECDC). The number of cases or deaths reported by any institution—including the ECDC, the WHO, Johns Hopkins and others—on a given day does not necessarily represent the actual number on that date.

Tests: we aggregate testing data ourselves by collecting it directly from national governments and health authorities. This allows us to adjust the full time series if it is published by the source.

sjshamji commented 3 years ago

Thanks for the data, it's awesome. I noticed a few other negative values. Not sure if this helps but in case it does, here they are:

Screen Shot 2020-11-04 at 3 36 13 PM
George6328 commented 3 years ago

Thank you for the data. I am comparing the ECDC and GitHub data on total deaths and confirmed cases vs daily and weekly intervals as well as the US CDC weekly data. One data set that is missing from JHU, ECDC, and GitHub is the age group death rate vs time. Only the CDC publishes this dataset. I find it disingenuous the media amplifying the total cases while the death rate has flattened. Additionally, the 65+ age group has become the dominant COVID-19 death sentence (especially with preexisting maladies). Can you provide me with any other datasets with respect to COVID-19 and age similar to the CDC?

George Gasparian

edomt commented 3 years ago

Hi @George6328 It's almost impossible to calculate a perfectly reliable 'death rate' at the moment, for many reasons but particularly the following ones:

I don't know of many datasets that provide age-specific data as we don't aggregate this data on OWID, but off the top of my head I can think of the French dataset here: https://www.data.gouv.fr/fr/datasets/r/08c18e08-6780-452d-9b8c-ae244ad529b3. The relevant dataset is the third one in the list, named donnees-hospitalieres-classe-age-covid19-YYYY-MM-DD-19h00.csv (reg is the area number so you can sum up all rows for a given day to get the national count, and cl_age90 is the age group with 0 as the sum of all age groups).

George6328 commented 3 years ago

Dear Edouard, Thank you very much for your information. I will review the French data. Sincerely, George

George Gasparian Sales/Applications Engineer​ GPD Optoelectronics Corp 7 Manor Parkway | Salem, NH | 03079

Visit us at booth no. at SPIE BIOS on 6-7 March 2021 at the Moscone Center, San Francisco, CA. Visit us at booth no. at SPIE Photonics West on 9-11 March 2021 at Moscone Center, San Francisco, CA.

From: Edouard Mathieu notifications@github.com Sent: Thursday, November 5, 2020 12:13 AM To: owid/covid-19-data covid-19-data@noreply.github.com Cc: George Gasparian ggasparian@gpd-ir.com; Mention mention@noreply.github.com Subject: Re: [owid/covid-19-data] Negative cases and deaths (#66)

Hi @George6328https://github.com/George6328 It's almost impossible to calculate a perfectly reliable 'death rate' at the moment, for many reasons but particularly the following ones:

I don't know of many datasets that provide age-specific data as we don't aggregate this data on OWID, but off the top of my head I can think of the French dataset here: https://www.data.gouv.fr/fr/datasets/r/08c18e08-6780-452d-9b8c-ae244ad529b3 (reg is the area number, you can sum them all up for a given day to get the national count, and cl_age90 is the age group with 0 as the sum of all age groups).

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/owid/covid-19-data/issues/66#issuecomment-722145940, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AQQ32PW2PVYRS4SO4TK3ERDSOIX6RANCNFSM4NMS3AAA.