This PR merges in the sequence notifications extracted by @dwolffram and joins them with the RKI case notifications.
@jbracher / @dwolffram could you sense check this all please? @dwolffram if you could update your extracted data again that would be great in order to catch the tail-end of Delta taking over in Germany.
The last plot in particular (here for convenience) is interesting as appears to show some issues with sampling for sequences that are reported early. Any ideas?I would say including this in the analysis vs excluding as seems like a feature of the dataset but open for discussion as will of course introduce bias in forecasts.
This PR merges in the sequence notifications extracted by @dwolffram and joins them with the RKI case notifications.
@jbracher / @dwolffram could you sense check this all please? @dwolffram if you could update your extracted data again that would be great in order to catch the tail-end of Delta taking over in Germany.
Markdown summary of data cleaning: https://github.com/epiforecasts/evaluate-delta-for-forecasting/blob/8c3111dde4fc84292f081558434d6cbfe9cfb25e/data-raw/observations/germany.md
The last plot in particular (here for convenience) is interesting as appears to show some issues with sampling for sequences that are reported early. Any ideas?I would say including this in the analysis vs excluding as seems like a feature of the dataset but open for discussion as will of course introduce bias in forecasts.