Public repository for CMU Delphi forecasting efforts for SARS-Cov2
-
forecasters: R packaged versions of forecasters; follow README here for instructions on how to prepare your forecaster
-
forecasts: Weekly forecasts merged from Monday production branch
-
utilities: packages used for production runs, evaluation, and other pipeline utilities for the forecasters
-
data: cached upstream data, data generated by training runs
-
docs: documentation for forecasters, utility packages
-
logs: training run logs
- Alden Green (Carnegie Mellon University)
- Addison Hu (Carnegie Mellon University)
- Maria Jahja (Carnegie Mellon University)
- Valerie Ventura (Carnegie Mellon University)
- Larry Wasserman (Carnegie Mellon University)
- Rob Tibshirani (Stanford University)
- Vishnu Shankar (Stanford University)
- Jacob Bien (University of Southern California)
- Logan Brooks (Carnegie Mellon University)
- Balasubramanian Narasimhan (Stanford University)
- Samyak Rajanala (Stanford University)
- Aaron Rumack (Carnegie Mellon University)
- Noah Simon (University of Washington)
- James Sharpnack (UC Davis)
- Daniel McDonald (University of British Columbia)
- Ryan Tibshirani (Carnegie Mellon University) and the Delphi COVID-19 Response Team
When adding a new forecaster or other package, take the following steps BEFORE you issue a pull request.
- On main/.github, copy one of the existing actions.
- Modify it to apply to your new package. And rename it following the conventions.
- Switch to develop and merge main onto develop to get your new Action there.
- Now switch to your package branch, issue a pull request to develop, and tag @dajmcdon or @jsharpna as reviewers.
If the check fails, revisit your package. Run R CMD Check locally and fix the problems. Make sure any dependencies are included as necessary. See the DESCRIPTION for {aardvark} as an example. If that all fails, get help on Slack.