These changes update my original OC immunity estimation model to use the CDPH data directly and adds a new ImmuneCohort model to estimate county-wide immunity more accurately. This new model is more sophisticated than the previous one but still results in a reasonably crude estimate based on a number of debatable assumptions.
Latest data and graphs can be found in this Google spreadsheet:
CDPH counts partially_vaccinated (PV) and fully_vaccinated (FV) independently. CDPH does not explain how it converts a PV (first short) to FV (second shot) record. Model assumes, when second shot is reported, CDPH decrements PV and increments FV counts on the date of the first shot.
Model does not try to filter out breakthrough cases from infected/recovered numbers. These will be double-counted for a while and may lead to overestimates (especially with the Omicron surge).
I would guess that the model overestimates immunity, perhaps significantly. It is also worth noting that immunity really exists on a continuum and includes critical factors like behavior not included here.
Screenshots
The following screenshots show the difference in estimates between the new model and the original:
These changes update my original OC immunity estimation model to use the CDPH data directly and adds a new
ImmuneCohortmodel to estimate county-wide immunity more accurately. This new model is more sophisticated than the previous one but still results in a reasonably crude estimate based on a number of debatable assumptions.Latest data and graphs can be found in this Google spreadsheet:
Notes and Assumptions
ImmuneCohortmodel constants here: models/oc/immune_cohort.pypartially_vaccinated(PV) andfully_vaccinated(FV) independently. CDPH does not explain how it converts a PV (first short) to FV (second shot) record. Model assumes, when second shot is reported, CDPH decrements PV and increments FV counts on the date of the first shot.Screenshots
The following screenshots show the difference in estimates between the new model and the original:
New
Old