Pollenomics Engine Roadmap¶
The current repository is an atlas-builder with reproducible evidence routing. That is the right starting point, but it is not the end state. The end state is a pollenomics runtime that can compare pollen, archaeological context, and ancient DNA evidence without losing provenance or contract clarity.
Current Stage¶
Today the repository can:
- collect tracked evidence from AADR metadata, LandClim, Neotoma, SEAD, RAÄ, and fieldwork records
- normalize those sources into reviewed files under
data/ - publish country bundles and one shared Nordic atlas under
docs/report/ - emit heuristic candidate-site ranking sidecars from the tracked atlas context layers
Next Stage¶
The next stage is to turn heuristic publication outputs into explicit workflow stages:
- candidate ranking should remain traceable to exact tracked context layers
- multi-evidence joins should state where provenance differs rather than flattening those differences into one score
- workflow outputs should be publishable and reviewable from the repository alone
Paper Direction¶
The planned paper direction is:
POLLENOMIC's: Decoding the farming history of Scandinavia using advanced statistics to combine ancient DNA with paleo-pollen data.
That future work needs more than a map. It needs reproducible statistical workflows, explicit evidence contracts, and a runtime that can relate pollen signals to ancient-DNA context without pretending that proximity alone is enough.
Boundary¶
This roadmap should keep ambition honest. A page like this is useful only if it prevents the repository from claiming that future engine work already exists in the current atlas-builder surface.