Getting Started¶
Installation¶
There exist three different possibilities to run the models:
Clone the repository, with the latest release:
git clone --branch v0.1.8 https://github.com/Priesemann-Group/covid19_inference
Install the module via pip
pip install git+https://github.com/Priesemann-Group/covid19_inference.git@v0.1.8
3. Run the notebooks directly in Google Colab. At the top of the notebooks files there should be a symbol which opens them directly in a Google Colab instance.
First Steps¶
To get started, we recommend to look at one of the currently two example notebooks:
- SIR model with one german state
This model is similar to the one discussed in our paper: Inferring COVID-19 spreading rates and potential change points for case number forecasts. The difference is that the delay between infection and report is now lognormal distributed and not fixed.
- Hierarchical model of the German states
This builds a hierarchical Bayesian model of the states of Germany. Caution, seems to be currently broken!
We can for example recommend the following articles about Bayesian modeling:
As a introduction to Bayesian statistics and the python package (PyMC3) that we use: https://docs.pymc.io/notebooks/api_quickstart.html
This is a good post about hierarchical Bayesian models in general: https://statmodeling.stat.columbia.edu/2014/01/21/everything-need-know-bayesian-statistics-learned-eight-schools/