Closed kaitejohnson closed 2 months ago
Should probably use the zoltr
API https://reichlab.io/zoltr/
Would you suggest 18/22 and 50/52? Fine with that, I don't think its likely to change the models chosen!
Would you suggest 18/22 and 50/52? Fine with that, I don't think its likely to change the models chosen!
I would suggest something more like 40/52. 18/22 seems good.
A little scope creep to get everything working needed to do the filtering to log scores. I am going to work off of this branch to remove the outlier wastewater data
This PR closes #114. It just expands the Hub models that we score (I had prototyped on a smaller set). These are all the ones highlighted for comparison here). However, there are a number of models (6 teams) that don't have forecasts for every date. We can continue to exclude them, or we could do something like allow one or two missing weeks.