This repository contains datasets that track the public reporting of Swiss COVID-19 data by the Federal Office of Public Health (FOPH) over time. It is based on snapshots of the public data from the FOPH COVID-19 dashboard.
The datasets show how many cases that occurred on a certain day had already been reported on the same day, one day later, two days later, and so on. This is useful for empirical analyses of reporting delays and for nowcasting of epidemiological data. The data is available both for Switzerland as a whole and for each canton individually.
Each dataset contains the following fields:
date_XThe date of occurrence of the corresponding event, where X is either case, hospitalizations, or death.date_reportThe date of public reporting by FOPH.geoRegionThe code for the respective canton or all of Switzerland, see here.totalThe cumulative number of events ondate_Xthat have so far been reported untildate_report. Ideally, this number should be strictly increasing over time. This is however not guaranteed, see Caveats.
Note that to reduce the size of the datasets, we only include reporting dates on which the entry for a specific event date was updated. For example, if the case count for a specific date changes on Friday and remains unchanged over the weekend, then Saturday and Sunday are not shown in the dataset, as they are identical to the entry on Friday. They can be easily reconstructed of course.
The recording starts at 2020-11-05 (first available snapshot). The time series resolution is daily, but updates are not published by FOPH every day: Until 2022-04-01, a daily reporting schedule except for weekends was used. After that, a weekly reporting schedule was used, with the new report being published every Tuesday.
Please be aware that since the publicly reported data is published by FOPH in an aggregated format, all corrections made to the data are subsumed in the cumulative count for a certain day. For example, if 100 new cases were reported for a certain day, but the dates of 10 already known cases for that day were changed to a different day as part of a correction, then the overall count will only increase by 90 cases for that day. Thus, the reporting delays as recorded by the aggregated data provided here may be slightly biased by data corrections. In extreme cases, this may even lead to decreases of the cumulative count, namely when there are more corrections than newly reported cases.
A particular anomaly to be aware of is in the hospitalization data from 2021-09-28 to (including) 2021-10-13. Here, the cumulative count dropped significantly for many hospitalizations dates. The counts were corrected again on 2021-10-14. It is therefore recommended to treat these data as missing and impute between 2021-09-27 and 2021-10-14 using some appropriate method.
The data in this repository was collated from the FOPH COVID-19 dashboard via the corresponding public API and is further distributed under CC BY 4.0. I take no responsibility for any usage of the data. The datasets will be updated in the future, but without guarantee of timeliness or completeness.
If you have questions about the data or their epidemiological usage, you can contact me via email.