Data sources for charts

This web page gives the sources of the databases used to generate the charts at the interactive webpage: https://www.iancampbell.co.uk/covid-19.php.


Some authorities publish cases data broken down by date of testing as well as by date of report (e.g. Public Health England for the UK). Where there is a choice, charts are drawn by date of report. This means that there are more ups and downs than showing by date of testing (due to the variable reporting delays), but it does also mean that all the data is immediately available for comparison instead of having to ignore the data from the last 5 or more days because it is known to be incomplete.

UK comparisons

Data for UK regional and local authority comparisons comes from the UK Government website [1].

At the start of the epidemic, information on the location of cases was given on all new cases, but from April to July, it was available on only around 25% of cases, because 75% of cases were diagnosed by the privatised testing sector read more. Since July, location has been available on all cases.

Scottish health boards

Data comes from the Scottish government website [2].

"Up until 14 June, the figures only include data from NHS Scotland laboratories. From 15 June the figures also include people tested through the UK Government's testing programme including drive through, mobile testing units and home tests. The increase on 15 June is due to the addition of all the UKG data on that day." This explanation is taken from the daily spreadsheet [3].

International comparisons

Data for country comparisons comes from the Johns Hopkins Hospital website:
  • the daily updates [4],
  • the spreadsheet of deaths [5], and
  • the spreadsheet of cases [6].

Data editing
The epidemics have put great strain on medical and public health resources, so it is surprising that the databases contain some missing values - i.e. there are some instances where the total for one day is identical to that for the following day, even though there are significant numbers of cases daily. It is clear that the updated figures have somehow not reached the database through delays of some kind. Such missing values have been dealt with by interpolation. This affects a very small proportion of the database values - about 20 values in total, and does not affect any conclusion from international comparisons.

China update
China announced on 16 April 2020 a revision of the number of deaths in Wuhan amounting to an increase of 1290 deaths [7], and an increase in total deaths in China from 3346 to 4636 - i.e. by a factor of 1.3855. This was dealt with by scaling up the numbers of deaths prior to 16 April 2020 by a factor of 1.3855. This does not appreciably affect any conclusion from international comparisons since the differences between countries are so large.

UK update
The UK revised its method for counting deaths in August 2020, so that deaths are counted if within 28 days of a positive test [8]. Figures shown on charts are the updated figures.

Spain
Figures collected by the Johns Hopkins University for deaths in Spain have shown some discontinuities; the worst one is that zero deaths were shown for 11 consecutive days from 8 June, followed by 179 deaths on 19 June. This was dealt with be replacement of the Johns Hopkins figures with figures from the website Worldometer [9]. This gives a Spanish Government website [10] as the source of the figures.


References

[1]Coronavirus (COVID-19) in the UK https://coronavirus.data.gov.uk/
[2]Coronavirus (COVID-19): daily data for Scotland https://www.gov.scot/publications/coronavirus-covid-19-daily-data-for-scotland/
[3]https://www.gov.scot/binaries/content/documents/govscot/publications/statistics/2020/04/coronavirus-covid-19-trends-in-daily-data/documents/covid-19-data-by-nhs-board/covid-19-data-by-nhs-board/govscot%3Adocument/COVID-19%2Bdata%2Bby%2BNHS%2BBoard%2B29%2BJune%2B2020.xlsx?forceDownload=true
[4]Johns Hopkins Coronavirus Resource Center https://coronavirus.jhu.edu/map.html
[5]John Hopkins Hospital spreadsheet of deaths https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv
[6]John Hopkins Hospital spreadsheet of cases https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv
[7]Reuters (17 Apr 2020) China says nearly 1,300 virus deaths not counted in Wuhan, cites early lapses https://www.reuters.com/article/us-health-coronavirus-china-toll/china-says-nearly-1300-virus-deaths-not-counted-in-wuhan-cites-early-lapses-idUSKBN21Z052
[8]https://www.bmj.com/content/370/bmj.m3220
[9]https://www.worldometers.info/coronavirus/country/spain/ accessed 27 June 2020

[10]https://www.mscbs.gob.es/profesionales/saludPublica/ccayes/alertasActual/nCov-China/documentos/Actualizacion_148_COVID-19.pdf


First published: 27 Mar 2020
Last updated: 4 Nov 2020