The Grand Ideas Issues

[pearsonca]

Use case

This issue will serve as a parent for interesting ideas for analyses, alternative data streams, model extensions, etc

As we (and others) identify these opportunities, we would like to collect them as sub issues here for the time being.

Proposed solution

No response

Alternatives considered

No response

[pearsonca]

More Observations Types: Incidence and Hybrid

Currently, the model works only on prevalence snapshots.

However, surveys like ChildVaxView have an explicit incidence component: there's a recruited group, which has a prevalence upon recruitment, and then longitudinally reported changes in prevalence -- aka, incidence. We are currently handling those snapshots as independent looks at prevalence.

[pearsonca]

More Flexible Hierarchy

Currently we force a particular number of levels (3), when more or fewer may be appropriate to the data. We also don't really have a way to support "ragged" hierarchy (e.g. some intermediate level(s) with no internal subgroups) or incompletely observed partitions (e.g. cases where we know the existence / size of an unobserved population, but have no direct observations).