Unified SIRNet Codebase and New Live Dashboard

This release brings together previously segregated modules. Data retrieval is automatic and well-parameterized. SIRNet can fit using either a linear or LSTM model for the infection rate at the country, state, or county level. A nice argparse interface allows for the specification of all command line options. A live dashboard of SIRNet forecasts and mobility data for the state of texas is available here.

SIRNet Integration with SEIR

SIRNet now supports SEIR modeling and many other considerations. Plotting is enhanced and easy generalizability is on the way.

Release v0.3.0 - now with SIRNet

New results and projections are available now - we forecast for a variety of regions and exploit mobility data in a novel hybrid SIR-neural architecture to forecast at varying granularity. Forecasts are given at the county- and state-level in the US and country-level worldwide.

Now with Bexar County Forecasting

We have added a new notebook for forecasting COVID-19 in Bexar County, TX, USA. The model has been fine-tuned some more and more data is being integrated. Stay tuned for further updates!

v0.1.0 Initial Release

This is the preliminary release of an experiment to forecast the growth of the COVID-19 disease in China. Expect new data, results, and analysis soon!