General approach to scenario analysis / post-processing?

[pearsonca]

I have another carry-over library effort that I think might be relevant to flepiMoP - {scenarios} - and I'd like to understand if it is indeed relevant.

As suggested by name, the package is indeed to aid in scenario analyses. Briefly, the idea is to provide a succinct-but-expressive language to capture what people want to do in scenario analysis comparisons. You provide:

1. what features define a scenario
2. which scenario(s) form baseline for comparison + what features to otherwise match by
3. what operation(s) define the scenario comparison (e.g. calculate cumulative effectiveness as a function of time)
4. optionally, what summaries to compute across samples (e.q. quantiles)

... then the package executes.

Currently, scenarios is written in terms of consuming data.table objects, but I see it moving the direction of creating a mini-language and then applying that across different storage / materialized formats.

What I'm interested in learning: what's going on in flepiMoP for this kind of generic scenario analysis framing? Or if not generic, some examples of what people have done expressing scenario analysis problems. Would be great to chat through some examples, and then see if {scenarios} can make problem / analysis expression easier.