Jump to content

Talk:COVID-19 pandemic by country and territory/Archive 8

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 5Archive 6Archive 7Archive 8Archive 9

Remove old table. Keep the table updated daily by bot

See also: Template talk:COVID-19 pandemic data/Archive 18#Replacement

The table in this section needs to be removed from the article:

It is being maintained mainly by 1 or 2 people lately (see history) vetting hundreds of sources. Here is the table template:

There is no way the above-linked table can be as accurate, or as regularly and fully updated, as this table updated daily by a bot:

It is found in this section:

Its reference: {{COVID-19 data/Cite}}:

  • Mathieu, Edouard; Ritchie, Hannah; Rodés-Guirao, Lucas; Appel, Cameron; Giattino, Charlie; Hasell, Joe; Macdonald, Bobbie; Dattani, Saloni; Beltekian, Diana; Ortiz-Ospina, Esteban; Roser, Max (2020–2024). "Coronavirus Pandemic (COVID-19)". Our World in Data. Retrieved 2024-11-17.

The link in that reference:

Then scroll down to "Download database" box and clicking it to go here and see databases:

--Timeshifter (talk) 11:52, 28 September 2021 (UTC)

Support: I agree that the original table has outlived its usefulness now that we have a reliable source that programmatically collects the same data. For the most part, the numbers look nearly 1:1 between the two tables and the "recovered" counts have been unreliable/useless for a long time.
It looks like the original table includes some territories that are missing from OWID. I think in most cases, OWID has the data for these regions. The entries just haven't been added to Wikipedia's OWID table.
Here's the full list (possible false positives for alternate spellings):
Abkhazia, American Samoa, Anguilla, Artsakh, Aruba, Bermuda, Bonaire, British Indian Ocean Territory, British Virgin Islands, Cape Verde, Cayman Islands, Curacao, Donetsk PR, East Timor, Falkland Islands, Faroe Islands, French Polynesia, Gibraltar, Greenland, Guam, Guernsey, Isle of Man, Jersey, Luhansk PR, Macau, Montserrat, New Caledonia, Northern Cyprus, Northern Mariana Islands, Puerto Rico, Saba, Sahrawi Arab DR, Saint Helena, Ascension and Tristan da Cunha, Saint Pierre and Miquelon, Sint Eustatius, Sint Maarten, Somaliland, South Ossetia, The Gambia, Transnistria, Turks and Caicos, U.S. Virgin Islands, Wallis and Futuna, Antarctica
- Wikmoz (talk) 03:04, 29 September 2021 (UTC)
Wikmoz and Tol. Hey Tol. I suggest adding all of the above subnational areas, territories, etc. to the 2 templates:
Template:COVID-19 pandemic death rates
Template:COVID-19 pandemic death rates by country
And any others you see in the database.
--Timeshifter (talk) 12:07, 29 September 2021 (UTC)
Hello! I've added the Freely Associated States — Federated States of Micronesia, Marshall Islands, and Palau — which have been added to OWID since I set up the bot. The list is a mix of other spellings, unrecognised states, and parts of other locations; OWID does not have data for any of them. Tol (talk | contribs) @ 14:32, 29 September 2021 (UTC)
Wikmoz and Tol. I checked the first 9 on the Wikmoz list. 6 of them are found on this cases page based on the WHO database:
COVID-19 pandemic cases#2021
I checked the cases page first because some of these smaller areas haven't reported any deaths.
I figured out only recently how to extract some info from that massive WHO .csv file in order to update the Sep 1, 2021 numbers on the deaths page here:
COVID-19 pandemic deaths#2021
See: Help:Table#Picking selected dates from massive .csv files
Tol, would it be possible to use the WHO database instead of OWID?
--Timeshifter (talk) 18:16, 29 September 2021 (UTC)
@Timeshifter: It would require overhauling most of the bot. More locations isn't necessarily better, it just implies a wider definition of what constitutes a country. Most missing locations have questionable sovereignty; nearly all have their numbers included in another location. I don't think it's an issue; it's just different thresholds of inclusion. Tol (talk | contribs) @ 18:51, 29 September 2021 (UTC)
Tol and Wikmoz. It is fairly common to include some territories and other subnational areas on country lists. But I don't see deaths per million in the WHO pages, tables, and databases I have looked at so far. I guess WHO just provides the raw numbers and lets groups of people from John Hopkins, OWID, etc. crunch the numbers further into rates, etc.. --Timeshifter (talk) 19:57, 29 September 2021 (UTC)
OWID is very deliberate in their country/territory selection so I'd go with whatever they have. Tol, if there are loud complaints, is it still possible to add a section to the bottom of the list for manual input as done previously for COFA in Template:COVID-19 vaccination data? - Wikmoz (talk) 04:27, 30 September 2021 (UTC)
@Wikmoz: In short, not really. It's technically possible, but brings problems of double-counting, would require another (fairly large) addition to the module, and would introduce the issue of templates having different numbers which this system was supposed to fix in the first place. On an entirely unrelated note (for @Timeshifter too), is everyone fine with moving this discussion to Template talk:COVID-19 pandemic data? I'd much rather prefer to avoid discussion forks. Tol (talk | contribs) @ 04:33, 30 September 2021 (UTC)
Tol and Wikmoz. I think we should keep this discussion here. The old template may get deleted, along with its talk page, when that template is no longer transcluded in any article. This is the only article it is transcluded in currently. A note could be left on that talk page pointing here for further discussion. That way nothing is lost.
Rather than try to paste an addition for more territories, etc. to this template, I suggest letting others create a separate template for that. It wouldn't have to be a bot. It could even be updated weekly or monthly. That template would go in a separate section here.
--Timeshifter (talk) 09:57, 30 September 2021 (UTC)
This may be a non-issue and should not be a blocker for moving forward. If it does come up, I think the first point of resolution would be to direct the user to OWID and request that they add the territory to the source data. As a fallback, it looks like there's an ugly hack that is possible (see bottom row of table) without modifying the module. - Wikmoz (talk) 17:41, 30 September 2021 (UTC)
I really don't think it should be a problem. I plan to switch the table over tomorrow. Tol (talk | contribs) @ 19:04, 30 September 2021 (UTC)
I noticed that a longer discussion is taking place here. Is it the content of Template:COVID-19 pandemic data or the template itself that will be replaced? It appears that Template:COVID-19 pandemic death rates would duplicate the replacement, with the only significant difference being that the death rate is also included. LSGH (talk) (contributions) 04:39, 2 October 2021 (UTC)

LSGH. Template:COVID-19 pandemic data and its section would be removed from this article. The section containing Template:COVID-19 pandemic death rates would remain. --Timeshifter (talk) 12:32, 2 October 2021 (UTC)

Timeshifter: Will it be possible to classify the template as a historical page? Some obsolete pages in project space are retained for historical reference. LSGH (talk) (contributions) 16:59, 2 October 2021 (UTC)
LSGH. I don't know. Maybe the Covid wikiproject people would know.
 Done. I've updated the template to use the module, and removed Template:COVID-19 pandemic data from the article. This way, in case someone uses the template, it's automatically updated. Tol (talk | contribs) @ 18:35, 2 October 2021 (UTC)

"The table is updated daily by a bot"

@Timeshifter: I would prefer to remove "The table is updated daily by a bot" (and links to Template:COVID-19 data/Cite) from COVID-19 pandemic by country and territory#Total cases, deaths, and death rates by country (Our World in Data) and COVID-19 pandemic by country and territory#Total cases, deaths, and death rates by country (Our World in Data). This is somewhat of a self-reference, and I don't think it should be in the text of the article. I do not think the link to Template:COVID-19 data/Cite is helpful; that template is just the citation (displayed in the reference in the table). A better alternative would be to use Template:COVID-19 data/Date for an automatically-updating date for when the data was last updated. Tol (talk | contribs) @ 16:59, 3 October 2021 (UTC)

Tol. I changed it to "The table is updated automatically daily." The note explains further how. The dating template is a good idea. I added it to both those sections just above the table.
Noting a map or table's update date, frequency, etc. is not a self reference. See the map captions noting that the date is found at the Commons source.
It is good to have most info though in notes or references. I find the 2 bot tables in this article to be poorly referenced. Some might even claim they fail verification because the reference is too general. So I would like to add additional reference material outside of the table templates.
The additional reference would be in the intro to the section containing the bot table. Note that there are multiple references for some of the maps and tables. What is the exact reference being used by the bot for the vaccine table? And for the death rates table?
I have had similar discussions concerning other tables. For example, see:
Note the detail of some of the references. It makes some of the references actually usable. Otherwise, we tell people to look around for an hour in a many-paged document. The reference here links to a general page with many links to many multi-paged documents. --Timeshifter (talk) 19:49, 3 October 2021 (UTC)
@Timeshifter: I think that's better. I think the reference is better than a reference to OWID's JSON file (which is the immediate source of the data), but the link to Template:COVID-19 data/Cite really doesn't provide any more information; it just restates the reference. I've changed it to Template:COVID-19 data, which provides a large table with most data. I think mentioning the bot in a footnote is fine. There really isn't a better URL, and all of the data are available (rather easily) from there. The reference for the whole table is right in its title. I strongly disagree with the first reference in the linked article containing "Please update the tables here only from this WPB source" — that should be in a comment/editnotice/something else, not in a reader-facing reference. The directions on how to access the data, though, are helpful in ensuring that they are verifiable. Do you think a direct link to the source ([1]) would be helpful? Tol (talk | contribs) @ 20:24, 3 October 2021 (UTC)
Tol. The reader facing note at List of countries by incarceration rate seems to be working better than the editnotice at List of countries by intentional homicide rate. We have had may discussions at both those pages about all of this over the years. The more notes and editnotices the better. We have less problem edits the more we add such info. It may not be pretty but it works better for the regular editors.
Some people want to see the expanded table you linked to. Some people want the detailed info and links in the info below the table. I will link to both in the note:
Template:COVID-19 data
Template:COVID-19 data/Cite
We need both the main reference, and the link to OWID's JSON file. See how I do it here:
Template:Monthly cumulative COVID-19 death totals by country 2021#cite note-WHO-csv-1
I give the link to the specific data file, and where it is linked from.
In the case of your templates the additional references may need to go in the article sections, and not the template. Since I believe you are using that main reference for multiple templates.
I don't know why people think less info is better. Wikipedia is not paper: WP:NOTPAPER. Some people think less is more with references. From my experience less is just less. And inadequate, and causes many problems that are easily avoided by just adding more info in the reference to begin with.
People should not have to trust the template maker, or the article editors. Wikipedia is based on verifiability, not trust. WP:V.
I want to be able to download the data file, and open it in freeware LibreOffice Calc, and check the numbers myself. For any list article or template.
--Timeshifter (talk) 22:00, 3 October 2021 (UTC)
@Timeshifter: I still maintain that the link to the citation is useless. The "detailed info and links" is at Template:COVID-19 data/doc; Template:COVID-19 data/Cite just has the citation again (and transcludes the documentation). As for number of references, there shouldn't be more than a few for any statement (maybe a few more for contentious ones) (Wikipedia:Citation overkill). For the tables, only one reference is needed. If you want to verify the data, just follow the link ([2]), click on a tab (say "deaths"), and you can download the data (in csv format, which is much better for spreadsheets) or view visualisations or get numbers. Tol (talk | contribs) @ 22:24, 3 October 2021 (UTC)
My first reply in the next section shows why it is important to have detailed referencing for tables. --Timeshifter (talk) 04:21, 4 October 2021 (UTC)

Concerns

Bit of a long one this. I’ve been recording the available coronavirus statistics on a daily basis since the 9th January 2020 and started noting down any figures I heard in media reports from China on a spreadsheet on 3st December 2019. Since then this has grown as statistical websites have become available to include daily country by country figures from Wikipedia, WorldOMeter, Testing, Vaccines, UK, England, Scotland, Wales, N. Ireland, Rep of Ireland and various UK regions into a 38Mb file collection of daily data.

To the spreadsheet above I add my own columns e.g. on the paste copy of the <https://en.wikipedia.org/wiki/COVID-19_pandemic_by_country_and_territory#Total_cases_and_deaths_(mixed_sources)> table that I take every day. I add new Death Rate, Daily Deaths, Daily Cases and Recoveries columns taking the previous days data from the latest data. By doing this every day on a spreadsheet it’s easy to cross check figures and I know what is updated and what is not. The Wikipedia table is usually the most accurate, bots or no bots. It’s very easy to see what is updated and what is not updated on a daily basis.

As a result of copying the figures on to a spreadsheet it’s easy to spot some of the inconsistencies and occasional errors on the reporting of daily figures on the UK countries websites, the Republic of Ireland’s government’s website and very occasionally on the Wikipedia table if data had been entered incorrectly, maybe an extra digit, wrong digit etc. If this happens I can cross check with the links for each country’s coronavirus reporting website provided in the notes on Wikipedia and correct any mistakes accordingly. The way it is since yesterday, users have no idea where this data is coming from as there are no references provided in the Wikipedia notes anymore. For instance, one example of these errors that occurs on an almost daily basis is that I noticed the cumulative figures for coronavirus deaths in certain English regions decreasing instead of increasing each day resulting in minus daily deaths. This is still happening in about 2 or 3 regions a day but back in March / April of this year there was a decrease in the total death figures of about twenty to forty deaths a day with no explanation given.

Just a few points.

1 – I don’t know about programming or about Bots but I do know about statistics and it’s usually garbage in, garbage out. I used to look at the OWID site last year but the data on it wasn’t updated as often as the Wikipedia or WorldOmeter sites and sometimes some data wasn’t updated for weeks or was inaccurate. This may have changed since then and it looks as if there is a lot more data sets than there was.

I’ve just done a quick check for today’s UK figures which are posted once daily and it looks as though nothing has changed in the past year on OWID. OWID is still out of date and not only out of date, it is also posting incorrect figures. For example, the UK Deaths for two days ago was 127, yesterday was 121, today is 43. <https://coronavirus.data.gov.uk/details/deaths>

OWID latest UK deaths is showing 124 ( wherever they get that from, there has never been 124 deaths reported on any single day since the beginning of the pandemic in the UK). <https://ourworldindata.org/covid-deaths>

Today’s current official UK Total Cumulative Death Total within 28 days of a positive test is 136,953 as at the top of the page on the .gov.uk link above. Wikipedia now shows this as 137,295, an increase of 382 deaths since yesterday’s correct Wikipedia total. There are 43 new deaths in the UK today, not 124 or 382. OWID is both inaccurate and out of date. This Wikipedia table is now a nonsense. I wish editors who don’t know what they’re doing would just leave well enough alone. If it’s working, leave it. If the figures are not accurate as is the case now, you’ll soon get to know about it. If it is to be changed, then someone who knows what they’re doing needs to change it and the automated update needs to scrape all the government sites around the world as in the reference notes for the original table for the official figures i.e. as in the table until yesterday.

Today’s official UK case cumulative total is 7,900,680. <https://coronavirus.data.gov.uk/details/cases>

Wikipedia / OWID case cumulative total is 7,908,091 or as in the link below the figure is listed as ‘7.91 million’. Accurate it is not. <https://ourworldindata.org/covid-cases>

Some countries update in real time and so their figures do update throughout the day depending when you check them but the UK is not one of them. They release their new official figures every day 16:00 pm. There’s no argument here, the Wikipedia / OWID figures are now inaccurate and should no longer be relied on. They’re about as ‘accurate’ as they were last year. If you check the links above you can see this for yourself.

I could go on but I hope it’s point taken.

2 – There used to be a noticeable Yellow Box Note at the top of the ‘Talk:COVID-19 pandemic by country and territory’ page last year, <https://en.wikipedia.org/wiki/Talk:COVID-19_pandemic_by_country_and_territory> with a warning not to use the World In Data statistics, I think it was, in Wikipedia as they were unreliable. I agreed with this then and after checking some of the figures today, I still agree with it although I see that the warning box has been removed. There is a note about accuracy on WorldOMeter.info from April 2020 so it could be this that I’m thinking of. This may have been the case at the time but they appear to me to have been accurate for about a year now, more accurate than OWID anyway. I hope I’ve shown the inaccuracy of the OWID data above.

3 – Wikipedia / OWID now has 194 Countries and Territories covered, <https://en.wikipedia.org/wiki/COVID-19_pandemic_by_country_and_territory#Total_cases,_deaths,_and_death_rates_by_country_(Our_World_in_Data)>

Old Wikipedia table had 243 <https://en.wikipedia.org/wiki/COVID-19_pandemic_by_country_and_territory#Total_cases_and_deaths_(mixed_sources)>

WorldOMeter which is now more accurate than Wikipedia has 229 <https://www.worldometers.info/coronavirus/#countries>

Not now included - Abkhazia, American Samoa, Anguilla, Antarctica, Artsakh, Aruba, Bermuda, Bonaire, British Indian Ocean Territory, British Virgin Islands, Cayman Islands, Charles de Gaulle, Coral Princess, Costa Atlantica, Diamond Princess, Djibouti, Donetsk PR, Falkland Islands, Faroe Islands, French Polynesia, Gibraltar, Greenland, Greg Mortimer, Guam, Guernsey, HNLMS Dolfijn, Isle of Man, Jersey, Luhansk PR, Macau, Montserrat, MS Zaandam, New Caledonia, Nicaragua, Northern Cyprus, Northern Mariana Islands, Puerto Rico, Saba, Sahrawi Arab DR, Saint Helena, Ascension and Tristan da Cunha, Saint Pierre and Miquelon, SeaDream I, Sierra Leone, Sint Eustatius, Sint Maarten, Somaliland, South Ossetia, Transnistria, Turks and Caicos Islands, U.S. Virgin Islands, USS Theodore Roosevelt, Vatican City, Wallis and Futuna

From your list Cape Verde, East Timor and Gambia are included but well spotted, Curacao is not in the old Wikipedia, OWID or in WorldOMeter.info.

4 – ‘Even the few editors who remain are no longer able to update the template as often as they used to.’ I check these figures every day. I would say that the figures are updated more now than they were a year ago. There are also more countries being updated on an almost daily basis than were being updated last year.

5 – Leave ‘Country’, ‘Confirmed’, ‘Deaths’, ‘Recoveries’ columns in the order they are in and as it was, and add in a new ‘Deaths/Million’ column on the right if need be. More data is usually better than less data.

6 – ‘ Recoveries ‘ column data was updated on a daily basis on the <https://en.wikipedia.org/wiki/COVID-19_pandemic_by_country_and_territory#Total_cases_and_deaths_(mixed_sources)> table for most countries as I check it every day and new recovery figures appear against almost all countries on a daily basis, although twenty-one countries or territories out of two hundred and forty-two do not post any recovery data at all. This data gives a rough number of people infected in any particular territory over a two or three week period.

7 – I’ve had a discussion like this before back in June 2020 when a Wikipedia editor wanted to remove the world country by country coronavirus Testing statistics table from the main <https://en.wikipedia.org/wiki/2019-20_Wuhan_coronavirus_outbreak> page as it was then and deprive the Wikipedia world of these figures forever. This dismayed me so much that I had to open an account and comment. Thankfully he reversed his opinion and coronavirus Testing now has a page in it’s own right.

8 – Thank you to the contributors who keep these pages updated. Believe me I understand what a job it is if being done manually as it takes me about 45 mins to 1.5 hours each day (if there are errors to research and correct) and it does take a certain amount of dedication. A bot to do the work would obviously be an easier solution but it should be Wikipedia’s own bot with the tables updated from the links in the previous reference notes as in the figures from each government. There must be lots of people out there in Wikiworld who could throw together a perfectly accurate bespoke bot from the original reference notes for these Wikipedia pages. I wish I knew how to do it myself. Until that project is able to be completed then I would propose that no changes are made and that the table remains as it was up until a day ago. I have referenced many of these government reporting websites myself since last year when I have suspected an error and although sometimes it is bit of an educated guess for myself to get the correct figure because of the different languages and script, they are, no doubt, the most accurate source of information and I feel both comfortable and confident with the figures I’ve had to enter from time to time by referencing the Wikipedia reference notes for each country. When you’ve been doing this a while, it’s plain to see that many governments around the world have their own agendas but they are the best that we have at present. Although seemingly convenient, OWID is really not our solution and it’s just not good enough for Wikipedia.

9 - Trust in Wikipedia’s figures is important and I know that I for one have grown over time to trust the figures in Wikipedia’s table and the way that they have been compiled so far. That trust has dissipated overnight to the point that I know now that they’re useless. BringBackTheStats (talk) 02:26, 4 October 2021 (UTC)

@BringBackTheStats: I'll answer each of your points:
  1. Our World in Data is updated daily, and so is Wikipedia's data via the bot. The one-time change when transitioning from one system to another is inevitable; the bot-updated system does not even report daily deaths/cases. OWID sources case data from Johns Hopkins University. For the UK, JHU sources data from the UK government and the Scottish Government; the number should be higher because the numbers for Scotland are slightly more recent.
  2. I don't know what the consensus on OWID was in April 2020, but the current consensus is that they are reliable. The current consensus is also that WorldOMeter is unreliable.
  3. The number of locations does not correlate with accuracy. Here's a list of the locations you mentioned:
  4. Most countries were updated every 1–3 days, but updates were irregular and had to be done manually.
  5. Recoveries are not included as part of the data. If you would like to gain consensus for more columns, you can do that. It would have to be a different template, though, because this one needs to be narrow.
  6. I'm pretty sure recoveries were cumulative, and I don't see how this would give "a rough number of people infected in any particular territory over a two or three week period" — that sounds like active cases.
  7. "deprive the Wikipedia world of these figures forever": no, there are page histories so one can see previous revisions.
  8. Well, it is a Wikipedia bot (I run it), but setting up custom code for ~250 locations would be pointless when OWID conveniently collects them already. You can see a list of data sources here, it's nearly all ultimately from local governments.
  9. They're still reliably sourced, just different. Most of this difference comes from different methods, such as getting more recent data directly from more devolved governments.
— Tol (talk | contribs) @ 04:09, 4 October 2021 (UTC)
BringBackTheStats and Tol. Interesting. Why are the OWID and WHO total deaths for the UK so much lower than from the official UK site:
--
https://ourworldindata.org/covid-deaths
OWID csv and xlxs files opened in freeware LibreOffice Calc. Both have this number:
Oct 2, 2021: UK. Total deaths: 137,295
--
https://en.wikipedia.org/wiki/Template:Monthly_cumulative_COVID-19_death_totals_by_country_2021#cite_note-WHO-csv-1
WHO. Link is found in the Data Download section of the WHO Coronavirus (COVID-19) Dashboard:
https://covid19.who.int/info
https://covid19.who.int/WHO-COVID-19-global-data.csv
Oct 1, 2021: UK total (under "The United Kingdom"): 136,662
--
https://coronavirus.data.gov.uk/details/deaths
.gov.uk site: "Deaths with COVID-19 on the death certificate"
Oct 3, 2021: UK total: 159,716
--Timeshifter (talk) 04:19, 4 October 2021 (UTC)
@Timeshifter: This might have something to do with the UK government changing deaths to be all deaths occurring within 28 days of a positive test. JHU's notes (at "August 17: United Kingdom") say that they've "revised the historical death data to match this reporting". I'll try to look into thin further. Tol (talk | contribs) @ 14:18, 4 October 2021 (UTC)
Thanks Tol. I see the many changes in the UK data by searching for "United Kingdom" on that page. The Aug 17, 2020 change in question:
Government changes definition of death to those occuring within 28 days of a positive test. We have revised the historical death data to match this reporting. | time_series_covid19_deaths_global.csv | The change in definition results in a loss of around 5000 deaths from the official tally. | Data accessed from the official webpage on August 17 was used to recreate the time series file.
Another reason to trust secondary sources like John Hopkins University.
It is well known that the total death numbers worldwide are low since many people are dying outside of medical purview. So I am surprised that death certificates are not trusted. But then one might argue with the qualifications of coroners in various parts of the country in question.
I guess JHU has to make a call. This is a common problem on country pages, and why I much prefer country lists based on one source like World Prison Brief on incarceration rate lists. Otherwise, we have hundreds of editors (over time) arguing over data from hundreds of sources for hundreds of countries, territories, and other subnational areas.
Apparently, JHU and WHO are using the lower number based on "Deaths within 28 days of positive test by date of death". That number is also on the official gov.uk page:
https://coronavirus.data.gov.uk/details/deaths
--Timeshifter (talk) 16:54, 4 October 2021 (UTC)

@Tol: @Timeshifter: Hi. I don’t have the time to go through all your points and reply to you properly tonight but I have been looking into it a bit as you do. Forget about the UK figures. They’re in a world of their own. I could spend days explaining the bubble notes I’ve made on the UK spreadsheet when figures don’t look right, or figures have been revised up or down or sometimes just a mistake by someone. I might have a go at counting them all sometime just to see how many there are. There’s a good example of this in the NI coronavirus deaths registered which happened around the first or second week of 2021. It now looks as if no-one will ever correct it as I suspect it’s being transferred into the UK .gov website automatically otherwise someone would have picked it up by now.

Anyway, this isn’t about OWID not posting the correct UK figures and making up their own figures. That was just one example. I told you I could go on but I thought you’d get the point. Look at Ireland’s figures, look at France’s, look at Singapore, look at almost any country’s official figures. The OWID figures don’t bear any relation to any country that I have looked at. It could be the case that they’re actually a more accurate nonsense than the government websites but no-one will ever know. The officially compiled figures are all we have to go on. The OWID are like something they made up a while ago and compiled as they see fit with the aid of a bot merrily churning out figures that everyone seems to have forgotten about or at least that was the impression I was left with.

You said that OWID get their figures from John Hopkins so I spent an hour and a half on the John Hopkins website this evening trying to locate the definitive sources for their information and how their figures are compiled. I did searches on Google. I didn’t find a link on Google, so on to the John Hopkins. I expected to find loads of links and references all over their coronavirus homepage or at least in the footers saying, ‘ This is how we compile our data.’ Or ‘For a list of our sources please click here.’ I couldn’t find any such link or page. I ended up going down so many dead ends on that website that I became infuriated and gave up. I had expected to see a clearly marked page listing the references for each country like Wikipedia had. I couldn’t find it. I found a few references to data they had used for maybe four or five countries but a lot of those led nowhere, i.e. to private company apps that had given up the task or links to dead websites that hadn’t been updated since last year. Maybe I’m really dumb but I’ve been following these figures for a while now and have usually bee able to locate the source references. OWID are not the official figures. Wikipedia’s figures were the officially released figures for each country and it was excellent to get them all in one table. If someone does know of a link to their data sources for each country then please post it so I can have a look at it.

@Timeshifter : That is the correct UK deaths by date reported figure but not for the 1st Oct. It’s the figure released at 16:00 on the 30th Sept 2021. Those are the official UK figures on https://covid19.who.int/WHO-COVID-19-global-data.csv but the figures in the New Daily Cases column don’t look right. If you check on the Cases tab on the .Gov coronavirus website ( link above ) -> Cases by Date Reported -> UK Total -> Data you’ll see that they’re not the official figure. It’s something that the WHO is doing itself, i.e. adding a column that I’ve also added myself on my UK spreadsheet tab, calculating the difference between the previous day’s Cumulative Case Total figure from the present day’s release of the Cumulative Case Total as one day I thought to myself that those figures don’t add up. The official UK Daily New Case released figure is usually more than the figure in that WHO column and is currently about an average of about 600 more cases a day over the past month or so or to put it another way, the reported UK Cumulative Total Cases figure is about 18,000 cases short every month to the officially reported Daily New Cases at present but that fluctuates over time. It could just be that collectively as a nation the UK is just not very good at adding up but a few more teachers and a reduction in class sizes should rectify that swiftly.

You’re only scratching the surface but I’m glad that you’re slowly getting sucked into the nuanced, intricate and muddy world of coronavirus data manufacture. Now multiply all of this by the number of countries on this planet by all the various social pressures in our modern world and I hope you can see how complicated this is. Now I don’t have anything against bots, as I said myself I wish I had the knowledge to do it myself but I hope you can see why this is a world that is encouraged by blind trust in simple bots that once let off their leash should by rights entail just as much ongoing monitoring to ensure that they’re still on the right path but in reality are probably forgotten about at times and just let run their merry way. It’s a world that humans are needed to constantly monitor. I have confidence in the Wikipedia data because everything is referenced and I can check it myself and have done so many times. I don’t have confidence in the OWID data because I don’t know where it comes from, I can’t check it, it’s not referenced and their numbers that I have tried to check against official sources are something completely different from that official data. As I said before, until we know where these figures come from, they’re not credible and it’s a case of garbage in, garbage out. Again as I’ve said before this isn't about the UK figures, check the OWID data against other country's official data. BringBackTheStats (talk) 09:02, 5 October 2021 (UTC)

BringBackTheStats. Here is the JHU reference in the article here:
CSSEGISandData/COVID-19. GitHub COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University.
The README.md link there has all the data sources:
https://github.com/CSSEGISandData/COVID-19/blob/master/README.md
"DATA SOURCES: This list includes a complete list of all sources ever used in the data set, since January 21, 2010. Some sources listed here (e.g. ECDC, US CDC, BNO News) are not currently relied upon as a source of data."
The notes section of the article here lists the OWID sources, one of which is JHU:
COVID-19 pandemic by country and territory#Notes
Following those links takes one to the main OWID page for data sources:
https://github.com/owid/covid-19-data/blob/master/public/data/README.md
It has a table at the top that is very helpful. Scroll down further for much more info.
--Timeshifter (talk) 12:27, 5 October 2021 (UTC)

@Timeshifter: I’ll have to check links above again in more detail but I have checked some of the links already and I couldn’t find the data references when I looked yesterday. BringBackTheStats (talk) 03:49, 6 October 2021 (UTC)

BringBackTheStats. The first README file I linked to above is a huge list of JHU sources with direct links to the sources. Here is the link again:
https://github.com/CSSEGISandData/COVID-19/blob/master/README.md
The second README link is to the OWID sources. But you have to click further to get to the URLs for the sources. OWID uses JHU for cases and deaths. So that's easy.
For OWID vaccination sources, though, I should have linked to this:
https://github.com/owid/covid-19-data/tree/master/public/data/vaccinations
It says the sources are buried in locations.csv:
  • source_name: name of our source for data collection.
  • source_website: web location of our source. It can be a standard URL if numbers are consistently reported on a given page; otherwise it will be the source for the last data point.
I downloaded that file and it is too large to be opened by freeware LibreOffice Calc. I get this message: "The data could not be loaded completely because the maximum number of columns per sheet was exceeded."
Maybe you can contact OWID and ask them to put that part of the .csv in a separate file, or even better, on a web page.
https://ourworldindata.org/about#contact
If you find better access to the vaccination info sources please let me know. I will put it in the Wikipedia article references.
There is an OWID vaccination info FAQ:
https://ourworldindata.org/covid-vaccinations#frequently-asked-questions
--Timeshifter (talk) 16:27, 6 October 2021 (UTC)

Can you also show the number of recoveries in the COVID-19 table? Qhairun (talk) 15:36, 11 October 2021 (UTC)

Please I want the old COVID-19 statistic table back. Qhairun (talk) 15:41, 11 October 2021 (UTC)

@Qhairun: The new table does not include recoveries, and there is currently no way to include them. You can view the last revision of the table with recoveries here, or you can use JHU's website. Tol (talk | contribs) @ 18:26, 11 October 2021 (UTC)
@Tol: @Qhairun: You keep fobbing people off with these one line fob-offs containing links that you haven’t looked at yourself and appear to know absolutely nothing about, at least that’s how it looks to me. The links mean nothing if they don’t answer users’ concerns. The first link you have supplied is to a copy of the table the last time it had the correct official information on it. 12 day old data isn’t much good to anyone. Recoveries can easily be included by either 1) reverting back to the old correct data table that was continually updated by Wikipedia users or 2) adding them again to the new table.
The JHU website link you supplied is not a link to the recovery data that @Qhairun: is asking for. It’s as if you haven’t bothered to find it. To extract the JHU recovery data is a formatting ordeal in itself, when it used to be easily available to copy and easily read on the old table. Because of this, I’ve posted a small guide on how to do this below. I hope that it’s easy to follow. The JHU recovery data can be found here https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_recovered_global.csv on this page https://github.com/CSSEGISandData/COVID-19/tree/master/csse_covid_19_data/csse_covid_19_time_series but it’s not easy to get the actual data, at least not with the software that I possess.
JHU Recoveries Data Recovery Process
How to find JHU Github Recoveries Numbers on a JHU time_series_covid19_recovered_global.csv file using Libre Calc on a desktop.
I’ve tried this on my Excel but it wasn’t having it so I’ve had to install Libre Office to open these files.
1) Download and Open the .csv file from the first link above.
You may get the message,
‘The data could not be loaded completely because the maximum number of columns per sheet was exceeded.’
2) Ignore that and click ‘OK’ and it should open up okay. What you’ll see if it’s opened up okay, is was looks like computer code. Don’t be deterred, the steps below will show you how to change this into a semi-readable spreadsheet.
3) Expand Column A to roughly 13.5 cm and Centre Right to see various smaller Territories ( although not yet ) such as those of France, UK, etc.
4) Expand Column B to roughly 2cm and Centre Left to see various Countries but mainly to see Central African Republic..
5) Collapse columns C & D to 1cm or less as you don’t need them unless you want the latitude and longitude data for each region.
6) As Column headings are on Row 1066, hide rows 1 to 1064 as you don’t need them, it’s just computer code as far as I can tell.
7) Highlight columns E to XM and set column width to about 1.5cm so you can see the dates clearly on line 1066. Column XJ shows the data for the 12th Oct 2021 at the moment.
8) Freeze rows and columns at cell D1067 and we’re good to go with something that resembles a spreadsheet.
9) There is a slightly different but similar process to open up each of the other JHU .csv files.
10) If you’ve successfully been able to follow all of this to retrieve the JHU recovery data you will realise that the last time the JHU published any recovery data for most countries was on the 4th August 2021. There is data for a few territories up until the 6th August, after that date it’s lines of zeros.
11) The only place that I know of to get updated worldwide Recovery data was on the old Wikipedia table which was updated from each country’s respective dashboard. That data has now been taken away from us.
Sorry for the mix-up about latitude and longitude on columns C & D in point 5) which I had added to point 6) originally but that's the way it should be now.
BringBackTheStats (talk) 13:42, 14 October 2021 (UTC)
@BringBackTheStats: My mistake; JHU has stopped publishing recoveries data (see announcement). In that case, I advise directly using government sources. You could also obtain it by . Tol (talk | contribs) @ 19:36, 14 October 2021 (UTC)

Accessing the data

BringBackTheStats. See:

The Github csv links do not operate as they do in other places that I have found csv files.

There you have to click them once before downloading the "Raw" csv file. For example, go here:

Click on total_deaths_per_million.csv

It opens into a Github page with a scrolling window consisting of plain text csv. Download the "Raw" csv file. Open it in freeware LibreOffice Calc. It is a spreadsheet with the deaths per million data for every day for every country. --Timeshifter (talk) 00:41, 17 October 2021 (UTC)

@Timeshifter
I was talking specifically just about the JHU JHU.csv files. They're the ones with the formatting issues, not the OWID JHU .csv files. I haven’t found a problem opening up the OWID JHU.csv files on Github which so far open up like a normal spreadsheet for me. I think you're confusing JHU JHU.csv files with OWID JHU.csv files which I still maintain at this time are not quite the same thing. I'm still in the middle of checking to try and figure out what's going on. One thing’s for certain is that it can be really confusing. Another thing's a definite so far and that is that the JHU JHU.csv and the OWID JHU.csv files are not updating as promptly as the updating done by the Wikipedia users on the old Wikipedia table. Some of the number are out by 24 hours and some numbers have lagged by 48 hours. The discrepancy in the figures could be caused by an extra layer of time lag that’s added in when the figures are obtained from OWID, whereas previously the data was lifted direct from official government websites in most cases and almost updated instantaneously for some countries by Wikipedia users and then added to the table almost as soon as the new data was released, sometimes within minutes but usually within a few hours.
For the moment I’m sticking to just the plain data of total cases and total deaths. That’s enough for me at the moment. Yes, I see that downloading the ‘Raw’ JHU JHU time_series_covid19_deaths_global.csv file here https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv (from the link on the https://github.com/CSSEGISandData/COVID-19/tree/master/csse_covid_19_data page ) does avoid the reformatting issue with the JHU JHU.csv files and they also includes the correct figure for yesterday’s UK total deaths of 138,527. Today’s total is 138,584 which was released 4 and a half hours ago https://coronavirus.data.gov.uk/details/deaths . Then go to the 'Deaths within 28 days of positive test by date reported' heading, click the 'UK Total' circular check box on the right and then click on 'Data' in blue to the left. This would usually have been updated by now on the old table. From what you read below it looks as if this will update tomorrow morning for most people in Europe and Africa and late afternoon or evening for Asians depending upon where you are of course.
The new OWID Wikipedia table is showing a number of 138,940 deaths and 8,443,882 cases at the moment. As you can see these figures don't appear anywhere, not on the official .Gov.uk site or in any JHU JHU.csv data.
Sorting the reformatting issue is all great but unfortunately it has opened up a whole new line of JHU confusementology that up until now I was unaware of. The file above includes the UK total deaths but this latest JHU JHU dailyreports.csv file ( from the JHU csse_covid_19_data page – link above ) https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data/csse_covid_19_daily_reports/10-16-2021.csv does not. It gives the total individual numbers for each country in the UK i.e. England, Scotland, Wales and Northern Ireland but thankfully the numbers do match up. The date in the filename is the correct date for the data but the timestamp of the update within the file, if it’s US time, means that this JHU file is updated about 19 hours after the release of the UK data. That’s not how the old table worked. I really can’t see any reason how the bot table is an improvement on the old table. And this is just the first stage of this process, then it’s on to OWID and I’m still trying to get to the bottom of what goes on there with their data but once again thanks for the ‘Raw’ tip.
I also fail to see how all of this is up to me to check that the data is correct. That should have been done before making this change. I’m flagging it up as incorrect and surely it’s up to others to check. I’ve read through various Wikipedia pages by now and I haven’t seen a single request from other users that this change was needed.
BringBackTheStats (talk) 19:20, 17 October 2021 (UTC)

Update map legend?

For File:COVID-19 Outbreak World Map Total Deaths per Capita.svg there are no longer any countries with <0.1 deaths per 100,000 population, so it might be a good idea to remove the lightest shade in the legend. An update to the range of values for each shade might also be appropriate; there are now only 6 countries in the 0.1–0.6 range, 19 in the 0.6–3.3 range, 42 in the 3.3–18 range, 52 in the 18–100 range, 67 in the 100–555 range, and just 1 in the 555+ range. JACKINTHEBOX • TALK 10:33, 19 October 2021 (UTC)

@JackintheBox: That sounds like a good idea; you're welcome to do that. I think gaining consensus on Commons, where the file is, is probably more relevant. Tol (talk | contribs) @ 18:22, 26 October 2021 (UTC)

Tonga

Timeshifter, regarding your revert to my edit, from what i can tell, the table to which i added tonga currently does not appear to use citations, so i decided to not add a visible one to it, to conform with the entries already present. also, i am not sure how i would go about providing a citation for the table from which i removed tonga, since the removal of the entry results in there being no entry for which a citation could be provided. i had added a source in a comment so that other editors who questioned the edit could easily verify that the edit was warranted, and not have to look for one to confirm. because you reverted my edit, i am assuming that i am missing something here, but i cannot seem to figure out what it is. it has been a while since i last edited this page, so i apologize if i am unaware of any significant changes since then.

also, admittedly, i don't consider the age of my account to be evidence of knowledge of current editing standards. much of what i have learned is outdated now, and i still look to edits from younger accounts to understand what is now acceptable. similarly, i believe most of my edits on pages dealing directly with the pandemic were made during the first half of last year, so i acknowledge that much has changed on these pages since then. if there's anything you think i should have known, but did not appear to when i made my edit, i would be grateful if you could be more detailed in your explanation, because i admittedly could not understand your edit summary. thanks! dying (talk) 18:30, 29 October 2021 (UTC)

Dying. It looks like the table is using the Wikipedia Covid pages for those countries and subnational areas as references. For example:
COVID-19 pandemic in Tonga
So since that is apparently the accepted method up to now, I guess you don't need to leave a reference or hidden note here. I don't normally edit that table, and so I did not notice that the Wikipedia pages were the references. --Timeshifter (talk) 18:43, 29 October 2021 (UTC)
thanks for looking into that and letting me know, Timeshifter. in that case, would you mind either self-reverting your revert or reincorporating my edit in a manner that you see fit? this would allow us to avoid all semblance of an edit war, which i would greatly prefer. thanks. dying (talk) 19:16, 29 October 2021 (UTC)
Dying. OK. Good idea. Done! --Timeshifter (talk) 21:22, 29 October 2021 (UTC)
great. thanks, Timeshifter! dying (talk) 23:04, 29 October 2021 (UTC)

Svalbard

There have been two cases in Svalbard:

1. Oct 7 - "A sick Russian fisherman who was picked up at Bjørnøya on Wednesday night and transported to a hospital in Longyearbyen, has tested positive for Covid-19." [1]

2. Oct 19 - "A resident of Longyearbyen has been confirmed infected with Covid-19." [2]

Laniwov732 (talk) 01:08, 8 November 2021 (UTC)

@Laniwov732: I'm not sure, but Svalbard's cases may be included in Norway. Tol (talk | contribs) @ 02:28, 8 November 2021 (UTC)

References

  1. ^ https://lokalstyre.custompublish.com/pressemeldingnsyk-fisker-har-covid-19.6414690-321755.html
  2. ^ https://lokalstyre.custompublish.com/fastboende-smittet-av-covid-19.6416971-321755.html

The map 'Total confirmed cases by country'

Two countries (Kazakhstan in Central Asia and Slovakia in Central Europe) have crossed the one million mark for many days but their corresponding colours on the map have not been changed. Even many countries later crossed the one million mark have had their colours changed. — Preceding unsigned comment added by 101.78.179.28 (talk) 01:07, 14 December 2021 (UTC)

Please discuss here too:
File talk:COVID-19 Outbreak World Map.svg
The map authors may not be reading this talk page here.
--Timeshifter (talk) 09:11, 14 December 2021 (UTC)

Old number of cases' graphic

What happened to the old list that showed total number of confirmed cases from higher to lower (1st USA, 2nd India, 3rd Brazil, etc.) I don't want to see a random lists where Peru appears first, then jump to a random country with, let's say 1000 cases, then to other where there's been 200.000 or a number like that. It's just weird and unorganized MatsLP (talk) 10:17, 24 December 2021 (UTC)

You may find it weird, but it's not disorganized. The table (assuming you are talking about the one at § Total cases, deaths, and death rates by country (Our World in Data)) lists countries initially by number of deaths per million. To see the table sorted by confirmed cases, click on the column heading "Cases" (twice). — JohnFromPinckney (talk / edits) 11:41, 24 December 2021 (UTC)

Huge reductions in World and EU Deaths, Cases and Deaths/million ?

Moved to Template talk:COVID-19 data § Decreasing data
Tol (talk | contribs) @ 23:30, 27 December 2021 (UTC)

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Yesterday figures for 25th Dec 2021
Country Deaths/million Deaths Cases
World 684 5,392,790 278,891,403
European Union 1,994 891,943 52,820,293
United States 2,452 816,436 51,966,735
Today’s figures for 26th Dec 2021
Country Deaths/million Deaths Cases
World 612 (-72) 4,823,730 (-569,060) 237,803,829 (-41,087,574)
United States 2,452 816,463 (+27) 52,095,411 (+128,676)
European Union 1,590 (-404) 711,113 (-180,830) 41,158,525 (-11,661,768)

Does anyone know or can anyone explain why these OWID numbers have changed so much?

Has anyone found the reliable references for the information on the OWID table yet? I've been referred to the JHU references in the past but the OWID data is not the same as the JHU data so these can't be the correct references for OWID. Where do OWID get their numbers from? I can understand numbers remaining static if not updated over Christmas but they shouldn't reduce like this.BringBackTheStats (talk) 11:43, 27 December 2021 (UTC)

I see now that they've changed again to what they should be.BringBackTheStats (talk) 11:49, 27 December 2021 (UTC)
Sorry, my mistake, they've gone back to the Christmas day update figures. Looks like there's no update for the 26th. Strange that they would update on Christmas day but not on the 26th Dec. So, OWID can move pretty fast and change their updates if they want to. It's a shame they usually leave it 24-36 hours before they can get around to it.BringBackTheStats (talk) 12:04, 27 December 2021 (UTC)
I've noticed this and replied in § Death toll. Tol (talk | contribs) @ 23:15, 27 December 2021 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Death toll

The death toll has already crossed five million. It crossed five million many days ago. Please update!49.178.138.126 (talk) 07:39, 2 November 2021 (UTC)

How much more than five million is it? And how do you know? (That is, what is your source?) — JohnFromPinckney (talk / edits) 10:27, 2 November 2021 (UTC)
This is one source. You can also do your own research: https://www.straitstimes.com/world/as-covid-19-deaths-cross-5-million-hope-for-a-battered-world... Thanks.49.178.138.126 (talk) 13:02, 2 November 2021 (UTC)
This is one source that doesn't explain your alarm that the article is out of date. The article's first sentence begins, "The global toll in the Covid-19 pandemic was just shy of five million on Saturday (Oct 30)...", which doesn't explain why you said it crossed five million many days ago. Yes, I can do my own research, but you're the one coming with a claim; it should be because you have at least one usable source supporting it. — JohnFromPinckney (talk / edits) 16:18, 2 November 2021 (UTC)
@49.178.138.126: @JohnFromPinckney: The death toll passed 5,000,000 on the 31st Oct or 1st Nov depending where you are on this planet. Most of the data on this table is now out of date since it was changed to being updated by a bot on the 2nd October 2021. Please read my comments, arguments and links in the sections above. I tried my best to have it preserved in it’s original form ( as updated by Wikipedia users ) as the old table was a very good source of accurate Covid-19 information. The Wikipedia table used to be the most ‘as close to live’ Covid-19 table on the internet. Now the data is
1) between 24 and 36 hours out of date for most countries
2) not using official government data for a lot of countries.
3) the data is so unreliable that most of the data is almost unusable and I no longer reference it. It should be flagged as such beside the table on the main page.
4) virtually impossible to cross reference as Our World In Data reference Johns Hopkins University, but the OWID data produced in the table doesn't match Johns Hopkins University data.
WorldOmeter table here https://www.worldometers.info/coronavirus/#countries is now a more accurate table than this Wikipedia table, if you want a table that's more up to date.BringBackTheStats (talk) 17:01, 7 November 2021 (UTC)
I looked at the top entries of deaths per million (highest to lowest) in the worldometers.info table, and compared their numbers to deaths per million in COVID-19 pandemic by country and territory#Total cases, deaths, and death rates by country (Our World in Data).
The numbers are all very close to each other. Some higher, some lower. --Timeshifter (talk) 23:41, 7 November 2021 (UTC)
@User: Timeshifter: I really can’t think of what comparing one set of incorrect data with another set that is almost correct data will tell you. Some of the WorldOmeter data is correct, some of it is not.
I’ve tried and I’ve tried to explain this very carefully, posted no end of links about JHU and OWID data but no-one wants to check it. I’ve explained different reasons why the data on the new table figures is incorrect, unreliable, not to be trusted, out of date, lagging behind, unable to be referenced, etc, etc. I just don’t use it anymore and have had to reconfigure the data I had collected and now use other sources, the sort I can cross reference and check. Other users have also pointed out their various reasons why they object to this new table yet they are ignored, as I am, and the new table is persisted with. Where has all of this discussion about replacing the old table gone, along with the objections from other users? Is this how Wikipedia works? People object and their valid objections just get swept under the carpet. Are you able to bring back the previous discussion, links etc?
It’s really very simple. Most governments around the world have health care systems. I don’t know about the US as I’ve never been able to track down the US government’s Covid-19 dashboard. Every day these governments around the world issue their Covid-19 figures for the day on their ‘.gov’ websites or through their official press agencies. A few country’s data are a bit ambiguous but these are not the norm. Some are live or used to be, some issue death figures once a week, daily or whatever it is that each country does. These are the figures. No ifs or buts or ‘very close’ There’s one set of official figures issued for almost every country. These were the figures on the old table. They were exact as per the official numbers. These are not the figures on the new table. That’s it. BringBackTheStats (talk) 07:30, 8 November 2021 (UTC)
I have looked at JHU, OWID, WHO, and worldometers data. All their numbers for total cumulative deaths (and therefore deaths per million) are pretty close. Or as some of us say, "good enough for government purposes". I don't sweat the small differences.
See WP:TALK about talk page guidelines. 4 threads remaining on the active talk page is the default for the talk page bot. See User:MiszaBot/config, and example 2: minthreadsleft = 4
See the top of the talk page to find the archive links. You can search the archives too.
--Timeshifter (talk) 11:24, 8 November 2021 (UTC)
@User: Timeshifter: So you still refuse to check OWID with each country’s correct totals i.e. the accurate data?
You say you don’t sweat the small differences; that might have been okay for the old variants but now we have a variant that’s 4 to 5 times as infectious and the numbers are doubling about every two days. Up to date figures are VERY important now. It’s impossible to calculate anything with any accuracy as the OWID imput data is not accurate and as the case numbers are rising so fast. To add to that the OWID data is 1 to 1.5 days out of date depending where you are on the planet. Even more infectious variants may occur in the future with an ever decreasing factor of OWID accuracy.BringBackTheStats (talk) 11:13, 27 December 2021 (UTC)
I've copied this comment above back in here as this is where it should be as it is part of the 'Death Toll' discussion. It's not a comment on the 'December 2021: Omicron variant, decreasing data' discussion which was obviously some sort of glitch in the data. Can someone remove this comment from the purple section below and the other Decreasing Data Talk page?BringBackTheStats (talk) 13:46, 28 December 2021 (UTC)
BringBackTheStats, that was my mistake when I split the discussion and moved part of it. Sorry about that. I've removed the comment from below. Tol (talk | contribs) @ 20:38, 28 December 2021 (UTC)
@User: Tol:Thanks. The Wiki editing font isn't the clearest with it’s size, method, etc.BringBackTheStats (talk) 07:25, 29 December 2021 (UTC)
You're welcome. Tol (talk | contribs) @ 21:22, 29 December 2021 (UTC)

December 2021: Omicron variant, decreasing data

Moved to Template talk:COVID-19 data § Decreasing data
Tol (talk | contribs) @ 23:30, 27 December 2021 (UTC)

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Why are the death toll is decreasing? Isn't supposed to be more than 5 million right now? Can someone explain this? And also, if you can, please fix this. I'm so confused. I used to check the statistics everyday, but now since the numbers were unexpectedly decreasing, I decided to use Worldometer instead of Wikipedia. Qhairun (talk) 15:33, 27 December 2021 (UTC)

Qhairun, thanks for pointing out the incorrect death total. Tol maintains the table that updates automatically every day. There must be a bug. Current worldwide death toll in the table is listed as 4,260,075. That is definitely wrong. --Timeshifter (talk) 22:37, 27 December 2021 (UTC)
I've checked the source data and manually run the bot again; it's back to the correct number of 5400292 deaths. I'm not sure what's going on here; if the next day's bot update is wrong again, I'll set up a debug log to try and see if there's some time of day dependent problem with the source data. It also might have just been a temporary problem that has been resolved. Tol (talk | contribs) @ 23:13, 27 December 2021 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:COVID-19_pandemic_by_country_and_territory/Archive_8&oldid=1175852422"