rtcovidlive / rtlive-global

Globalized version of the Rt-live model.
Other
32 stars 13 forks source link
covid19 covid19-model nowcasting

Country-agnostic modeling of Rt

This repository contains code for data processing and modeling of COVID-19 reproduction number Rt using the model that was developed for https://rt.live.

While the implementation of the model is completely country-independent, the code has a high-level interface that allows for easy plug-in support of new countries.

The rt.live site itself (for the United States) runs the model hosted at rtcovidlive/covid-model, on which this further work was based. Other sites like rtlive.de run the code in this repository.

Where to find...

Adding Country Support

We learned that artifacts in data often require manual intervention to be fixed. At the same time, data loading and processing must be fully automated to support running the data processing for tens or hundreds of regions every day.

In this repository, we implemented a generalized data loading & processing interface that allows for:

Contributions to add/improve country support are very welcome!

Contributing

To contribute data sources, fix data outliers, or improve data quality for a specific country, please open a PR for the corresponding data_xy.py file in rtlive/sources.

Furthermore, we welcome contributions regarding...

How to run the code and notebooks

To be able to contribute, you'll need to be able to run the code and notebooks on this repo. To that end, we strongly recommend installing the miniconda distribution for Python. This will make sure all the packages and necessary compilers come from the same source, and it will allow you to create dedicated virtual environments to run your code safely for each project independently.

Once miniconda is installed, use the terminal and make sure you're in the root of the rtlive-global repository. Then, follow the following steps (still in the terminal):

See here for detailed instructions if you have issues.

Citing

To reference this project in a scientific article:

Kevin Systrom, Thomas Vladek and Mike Krieger. Rt.live (2020). GitHub repository, https://github.com/rtcovidlive/covid-model

or with the respective BibTeX entry:

@misc{rtlive2020,
  author = {Systrom, Kevin and Vladek, Thomas and Krieger, Mike},
  title = {Rt.live},
  year = {2020},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/rtcovidlive/covid-model}},
  commit = {...}
}