time-varying imputation of onset

[geryan]

swap out the ecdfs in notification_delay_cdfs for the ones in delays_from_onset?

Nick Golding 12 minutes ago
I think so

Nick Golding 12 minutes ago
might need a little plumbing

Gerry Ryan 11 minutes ago
cool - yeah it shouldn't be too tricky, he says confidently

Nick Golding 11 minutes ago
It’s possible it would be better with larger smoothing windows (so it is less reactive to little bumps), but just plumbing in this version would be a good start

Nick Golding 11 minutes ago
yeah, just pull out the most recent date for each state, and use the ecdf in the same way?