jinshuxian / covid19_metadata

Interoperable secondary data for research on the new COVID-19 coronavirus sars-cov-2

README

Caspar J. van Lissa 4/1/2020

COVID-19 Metadata

A collection of relevant country/city level metadata about the COVID-19 pandemic, made interoperable for secondary analysis. Curated by Data scientists Against Corona, collaborators: Caspar van Lissa, Tim Draws, Andrii Grygoryshyn, Konstantin Tomić, and Malte Lüken.

Available data sets

The following data sets have been processed:

Category	Information	Source	Folder	License	Reference
Response	Number of tests	Our world in data	OWID_Tests
Mobility	Google mobility data	Google	google_mobility
Risk level	Hospital data per country	WHO Health workforce/facilities database	WHO_OECD
Risk level	Health infrastructure per country data	OECD Health care resources database	WHO_OECD
Policies	Government effectiveness	Worldwide Governance Indicators	WB_GOV	CC-BY 3.0
Policies	COVID-19 specific regulation policies	Oxford Tracker for Regulation Policies	Ox_CGRT	CC-BY 4.0	Hale, Thomas and Samuel Webster (2020)
Preparedness	Global Health Security Index	Nuclear Threat Initiative	GHS	CC BY-NC-ND 4.0
COVID19	Number of cases and fatalities	CSSE Global Cases	CSSE	Copyright (academic use permitted)	Dong, Du, & Gardner, 2020
Economic	World Development Indicators	World Bank	WB_WDI	CC-BY 4.0

Folder structure:

Folder	Description	Permissions
data	Metadata sources in .csv format (intermediate formats are acceptable until they can be made tidy).	Do not edit
scripts	(R)-scripts	Human editable
doc	Documentation for your contribution, ideally in Rmarkdown format. Rmarkdown can contain code chunks. Elaborate functions should be relegated to the ‘scripts’ folder.	Human editable

How to use

Fork or clone this repository (for GitHub beginners: You can also click the green button that says “Clone or download”, and download a .zip). All data are in the /data folder. Some data are rarely updated (e.g., annual data), and some are updated daily. To ensure that you have access to the latest data for frequently updated sources, run the R-script in the run_me.R file, in the main folder.

Standards for data

Every source is condensed into one data file in .csv format, according to these specifications:

• Data should be available on the country- or community-and-country level.
• Recent data are the focus; if multi-year data is available, older years can be dropped
• All variable names should be lower case
• Mandatory variables are country (plain text country), and countryiso3 (ISO3 country code)
• Optionally, a region variable can be added
• Data should be in wide format: One row per country, one column per variable

Standards for data dictionary

A data_dictionary.csv is available for each data set, unless the file contents are immediately clear from the file. This data dictionary includes:

• variable: The name of the variable in the data file
• description: The description of this variable

Any other important information per variable can be included in this dictionary, such as sources, weights, etc.

News

The following issues are ongoing:

• Adding more databases; feel free to make a suggest or request a database here
• Added time-since first occurrence for Oxford policy / incidence trackers
• Added last observation carried forward for WHO data

License

This project is under a GNU GPL v3 open source license (see the LICENSE file). Individual data sources have different licenses; always check the license before publishing based on these data.

Contributing and Contact Information

This project is open for collaborators with valuable expertise. Contribute by:

• Filing a GitHub issue here
• Making a pull request here

By participating in this project, you agree to abide by the Contributor Code of Conduct v2.0.

A WORCS Project

This project is based on the Workflow for Open Reproducible Code in Science (WORCS). For more details, please read the preprint at https://osf.io/zcvbs/.

WORCS - steps to follow for each project

Study design phase

1. Create a new Private repository on github, copy the https:// link to clipboard
   The link should look something like https://github.com/yourname/yourrepo.git
2. In Rstudio, click File > New Project > New directory > WORCS Project Template
   1. Paste the GitHub Repository address in the textbox
   2. Keep the checkbox for renv checked if you want to document all dependencies (recommended)
   3. Select a preregistration template
3. Write the preregistration .Rmd
4. In the top-right corner of Rstudio, select the Git tab, select the checkboxes next to all files, and click the Commit button. Write an informative message for the commit, e.g., “Preregistration”, again click Commit, and then click the green Push arrow to send your commit to GitHub
5. Go to the GitHub repository for this project, and tag the Commit as a preregistration
6. Optional: Render the preregistration to PDF, and upload it to AsPredicted.org or OSF.io as an attachment
7. Optional: Add study Materials (to which you own the rights) to the repository. It is possible to solicit feedback (by opening a GitHub Issue) and acknowledge outside contributions (by accepting Pull requests)

Data analysis phase

8. Read the data into R, and document this procedure in prepare_data.R
9. Use open_data() or closed_data() to store the data
10. Write the manuscript in Manuscript.Rmd, using code chunks to perform the analyses.
11. Regularly commit your progress to the Git repository; ideally, after completing each small and clearly defined task. Use informative commit messages. Push the commits to GitHub.
12. Cite essential references with one at-symbol ([@essentialref2020]), and non-essential references with a double at-symbol ([@@nonessential2020]).

Submission phase

13. To save the state of the project library (all packages used), call renv::snapshot(). This updates the lockfile, renv.lock.
14. To render the paper with essential citations only for submission, change the line knit: worcs::cite_all to knit: worcs::cite_essential. Then, press the Knit button to generate a PDF

Publication phase

13. Make the GitHub repository public
14. Create an OSF project; although you may have already done this in Step 6.
15. Connect your GitHub repository to the OSF project
16. Add an Open Science statement to the manuscript, with a link to the OSF project
17. Optional: Publish preprint in a not-for-profit preprint repository such as PsyArchiv, and connect it to your existing OSF project
    • Check Sherpa Romeo to be sure that your intended outlet allows the publication of preprints; many journals do, nowadays - and if they do not, it is worth considering other outlets.