Pipeline stages

Notebooks for data processing and analysis (notebooks) are grouped into task types:

Stage	Description
01_configure	Define folder structure, set preferences, lock administrative referencing
02_ingest	Download, unzip, and stage data in cache
03_harmonize	Align entity datasets from multiple sources, create spine of entities
04_enrich	Derive attributes for entities using geoprocessing and record linkage.
05_curate	Create analysis-ready dataset (select, aggregate, snapshot).
06_model	Fit models, cross-validate, derive standard errors, optimize.
07_infer	Create artifacts from models: predictions, aggregated coefficients
08_report	Creation of publication-ready figures, tables, and text
09_show	Interactive display of results, demonstrating quality, issues, or functionality

These notebooks can be converted to scripts and be orchestrated across large computing clusters (to process data entire world regions).