Epidemiological data of COVID-19 for Denmark

This archive contains data from Statens Serum Institut (SSI), an institution under the auspices of the Danish Ministry of Health.

Most of the data in this repo has been extracted by me by scraping the official COVID-19 status web page of the Danish Ministry of Health. Although the the data is also available from the producer Statens Serum Insitut I found that the former web page changes far less frequently.

June 17, 2020 SSI changed the definition of COVID-19 related hospitalizations to be defined as hospitalization occurring within 14 days after the first positive test. Furthermore, patients still hospitalized 90 days after first positive tests are removed from the statistic. The numbers for hospitalizations (+ ICU, ventilators) have been updated accordingly.

Content CSV Data Files

covid19-data-denmark.csv

The data contained in the file covid19-data-denmark.csv includes the following columns:

day: days since Monday 2020-02-24, the first day of the week where the first COVID-19 cases were first recorded in Denmark.
date: The date in ISO 8601 format.
tests_accum: Accumulated number of performed tests. This data was not provided before 2020-05-15.
tests_today: Tests reported on this date. This data was not provided before 2020-05-16.
tested_accum: Accumulated number of tested persons by PCR tested one or more times. Daily updates to this data was not provided before 2020-03-02, the total of all tested persons before that (358) are recorded on this day.
tested_today: Persons PCR tested on this day that have not been tested before.
infected_accum: Accumulated number of confirmed cases by PCR test.
infected_today: Number of infected patients confirmed by PCR test on this date.
dead_accum: Accumulated number of deaths associated with the COVID-19 pandemic.
dead_today: Number of persons reported dead on this date.
recovered_accum: Accumulated number of recoveries after positive PCR result.
recovered_today: Persons reported recovered on this date after positive PCR result.
active: Active cases. Computed per day using this formula: infected_accum - recovered_accum - dead_accum
hospitalized_now: Number of patients currently hospitalized.
hospitalized_change: Change in hospitalizations since the day before.
hospitalized_admitted: Newly admitted patients. (*)
icu_now: Number of patients currently in Intensive Care Units.
icu_change: Change of ICU patients since the day before.
respirator_now: Number of patients currently in ventilators.
respirator_change: Change in number of patients in ventilators since the day before.

(*) The data in the hospitalized_admitted column is extracted from the dataset Newly_admitted_over_time.csv that can be found in a zip-archive updated on weekdays and published on SSIs web page.

All other data is scraped from SSI's COVID-19 dashboard.

Note1: 2021-03-10: The data from the date 2021-03-10 contains errors in the numbers. On this date, SSI began pooling antigen and PCR data from private test centres to the official ones, and apparently an error was made. This is extremely bad, since there are many consumers of the data published by SSI, and that data is now tainted, since we don't know how many patients appear two or more times in the statistics.

Note2: 2021-03-26. From this date, SSI has again revised the data reporting on their COVID-19 dashboard. Antigen tests and PCR tests are again separated. This means that data in the period 2021-03-10 (day 380) to 2021-03-25 (day 395) contained results from PCR + antigen tests, where many of the latter were false positives. Using various sources, including data scraped from this page, where PCR and Ag test numbers were given, the data in covid19-data-denmark.csv has been corrected so only PCR data is reported. The problem with antigen tests is that they are not conclusive, and patients are referred to an PCR tests to give the final answer. Corrections include the following:

infected_accum is now data from smitteapp.infected_accum for day >= 380 and day < 396. The column infected_today is computed from this as daily difference.
tests_accum contains data copied from the SST webpage as described above, and tests_today is computed from this as a daily difference.
The data in the columns recovered_accum and tested_accum is only reported on SSI's COVID-19 dashboard, so a correction procedure was carried out. Using the PCR-only values of days 379 and 396, the true increase could be computed, and the data was scaled proportionally to fit these constraints. This procedure has removed data from the antigen tests, and further the huge positive and negative spikes in the data resulting from the two changes in data reporting practice by SSI. However data in the columns recovered_accum, recovered_today, tested_accum and tested_today should be considered approximate in the range day 380 to day 395 as they are not the exact recorded numbers.

Note3: 2021-06-29. From this date, SSI no longer provides the data on first time tests, given in the fields tested_today and tested_accum.

vaccinations.csv

This file contains data of COVID-19 vaccinations, extracted from a PDF file published by SSI on weekdays.

day: days since Monday 2020-02-24, the first day of the week where the first COVID-19 cases were first recorded in Denmark.
date: The date in ISO 8601 format.
week: The ISO week number.
vaccination1_accum: Accumulated number of first vaccination
vaccination1_today: Today's number of first vaccination
vaccination2_accum: Accumulated number of second vaccination
vaccination2_today: Today's number of second vaccination

From the date 2021-02-22 SSI is publishing daily vaccination numbers on their vaccination dashboard, rather than from at table in a daily PDF file. It means yesterday's numbers are presented as today's; therefore there is no data for day 363 (2021-02-21).

private_testing.csv

This file contains data from private testing providers, extracted from a PDF file published by SSI on weekdays. It is not known how many of these patients also have been tested in the public testing system. Indeed it is quite likely that patients with positive antigen test have their results confirmed by taking a test in the public system, which is free. Therefore, the data in this table cannot be added to the data in the file covid19-data-denmark.csv.

day: days since Monday 2020-02-24, the first day of the week where the first COVID-19 cases were first recorded in Denmark.
date: The date in ISO 8601 format.
week: The ISO week number.
tests_pcr_accum: Performed PCR tests, accumulated
tests_pcr_today: Performed PCR tests, today
infected_pcr_accum: Infected patients found by PCR testing, accumulated
infected_pcr_today: Infected patients found by PCR testing, today
tests_ag_accum: Performed antigen tests, accumulated
tests_ag_today: Performed antigen tests, today
infected_ag_accum: Infected patients found by antigen testing, accumulated
infected_ag_today: Infected patients found by antigen testing, today

Note: This data is not published after 2021-03-10. After this date, the numbers implicitly appear in covid19-data-denmark.csv. Extremely poor data reporting practice, but what can we do?

mutant_data.csv

This file contains sequencing data on positive SARS-CoV-2 samples. The data is extracted from the statistics page of the Danish Covid-19 Genome Consortium. The data is aggregated per week.

day: The day number of the last day of the week (Sunday). Day zero is Monday 2020-02-24, the first day of the week where the first COVID-19 cases were first recorded in Denmark.
year: The year
week: Week number of the year.
variant: The genome variant.
ireg: The region number.
region: The region name
positive: Number of variant genomes sequenced
genomes: Number samples sequenced
cases: Total number of cases in that region and that week

Log of Registered Changes to the Source Web Page

Table of registered changes to the source web page. Only events that broke web scraping is recorded. The list of events before 2020-06-26 is not exhaustive.

Date	What
2020-04-02	Official recovery stats available
2020-04-10	Deal with date typo 10 .april
2020-05-16	More exhaustive data provided in Table 1.1. New fields in tables 1.1 and 1.2
2020-05-20	Section 1.1 title changed
2020-05-21	Table 1.2 changed
2020-05-12	SSI removes 6 wrongly categorised patients from deaths count
2020-05-25	Deal with date typo '24. ma' instead of '24. maj'
2020-06-08	Deal with date typo 'juní' instead of 'juni'
2020-06-18	Deal with typo in number of `recovered_accum` (12242 -> 11242) that made number of actives become negative
2020-06-20	Data is not updated on 2020-06-20 and 2020-06-21 due to server maintenance.
2020-06-22	Data collected during the weekend is reported on this date.
2020-06-26	Section 1.1, 2.4 and 3.7 titles changed.
2020-06-27	From this date, data is no longer updated on weekends.
2020-07-21	Section 3.7 title changed.
2020-08-24	Scrape data from [SSI's webpage] which now includes weekend data
2020-08-31	Section 3.7 title changed.
2020-10-26	Accumulated tests data for day 0 to 81 added from SSI's file `Test_pos_over_time.csv`.
2020-10-27	Between 2020-06-19 and 2020-08-24 data was reported collectively on mondays, weekend data in this range interpolated.
2021-01-12	Added data files `vaccinations.csv` and `private_testing.csv`
2021-02-08	Added data file `mutant_data.csv`
2021-02-22	SSI now published daily vaccination numbers on their vaccination dashboard
2021-03-10	From this date, SSI adds testing data from private test centres, antigen and PCR are pooled. This is very bad.
2021-03-25	SSI have again changed their data reporting, now antigen tests are no longer included, so this day's numbers become negative if you check. Extremely bad.
2021-03-26	Data has been corrected to exclude antigen test data, see notes 1 & 2 above

License

To the extent possible under law, Morten Kjeldgaard has waived all copyright and related or neighboring rights to Official COVID-19 data for Denmark. This work is published from: Denmark.

mok0/covid19-data-denmark