epiforecasts / covid.german.forecasts

Comparing crowd sourced and model derived forecasts of Covid-19 for Germany and Poland

https://cmmid-lshtm.shinyapps.io/crowd-forecast/

MIT License

4 stars 1 forks source link

Update robot #30

Closed seabbs closed 3 years ago

seabbs commented 3 years ago

Rt and secondary model forecast update. Finalised the secondary model to have a wider prior on the proportion of test positive cases that result in death and shortened the fitting window to 3 weeks. This is now locked in for evaluation.

seabbs commented 3 years ago

Currently seeing a cumulative forecast bug with all forecasts + only 3 forecast dates. Hmmm. Any ideas?

Screenshot 2020-12-14 at 11 48 00

seabbs commented 3 years ago

Ahh looking here there is a new year bug in the formatting code (as EpiWeeks rest at the beginning of the year).

seabbs commented 3 years ago

Looks like I've fixed it. @nikosbosse please check the code to see if it makes sense to you.

Rt forecasts

Screenshot 2020-12-14 at 12 35 36 Screenshot 2020-12-14 at 12 36 06

Convolution forecasts

Here struggling to track recent surge in deaths vs observed cases.

Screenshot 2020-12-14 at 12 37 09