Following up on the Covid-19 ForecastHub and the recent European hub, this note is to find collaborators for a worldwide hub. The idea is to work as an open research organization.

It is very hard to know how to compare and evaluate forecasting models that are made after the fact. There is value in open research on model forecasts.

Building on Our World in Data’s Covid-19 dataset and the GISAID hcov-19 variant data: https://www.gisaid.org/hcov19-variants/

Some longer-term extensions that I think would be particularly interesting: conditional forecasts, in particular conditioning on the RNA sequence of the dominant variant in different countries seems eminently useful.