Environmental analysis of COVID-19 transmission rates
Repository structure
/scripts/ contains bash and python scripts for downloading ERA5 atmospheric reanalysis data and COVID-19 clinical data.
/data/ contains processed data and is a placeholder directory that holds raw data files too large to be version-controlled.
/notebooks/ contains the key python notebooks used for pre-processing, analysis, and plotting.
Notebook naming conventions
We use the following prefix conventions for naming notebooks:
0_for pre-processing locally-downloaded raw data (large > 10 GB) and producing locally-saved interim data (10 MB < medium < 10 GB)1_for analysis of raw and interim data and producing plot-ready processed data (small < 10 Mb, pushed to github)2_post-processing and making publication-quality plots (saved in/figures/)T_testing and code development (avoid using inmasterbranch)
Programming environment
The python packages necessary for running the python scripts and jupyter notebooks included here are listed in the environment.yml file. We recommend using conda to install these packages using the command:
conda env update -f environmental.ymland activating the environment with
conda activate covid-weatherDownloading raw data
Downloading JHU CSSE COVID-19 epidemiological data
The /scripts/get_covid19_data.sh script clones the JHU CSSE dataset into the /data/ folder.
Downloading ERA5 reanalysis data
Our scripts use the Climate Data Store (CDS) API and require an account. These instructions describe how to configure your account key and use the python app, which is installed via pip install cdsapi.
Near-surface meterological variables in ERA5 (and quality-controlled bias corrected fields through 2018)
Some variables of interest:
| Variable | Units | API name for python script |
|---|---|---|
| Altitude | meters | grid_point_altitude |
| Temperature | Kelvin | near_surface_air_temperature |
| Specific humidity | kg water / kg air | near_surface_specific_humidity |
| Pressure | Pascals | surface_air_pressure |
| Rainfall | kg-meters^2 / s | rainfall_flux |
Instructions for computing the near-surface specific and relative humidities, which are not archived diagnostics, from the near-surface temperature, dew point temperature, and surface pressure.
Contributor guidelines
Contributors to this repository should use the following workflow to ease collaboration: 1) Fork the repository 2) Create a new branch with a name that reflects your intended contribution 3) Make local changes to your branch, following the repository structure and naming conventions 4) Add, commit, and push the local changes to your branch to your fork 5) Open a pull request and request review from a relevant co-contributor 6) Celebrate as your changes are approved by a reviewer and merged into the master branch!