| --- |
| license: mit |
| --- |
| # Microcosm cross-dataset evaluation bundles |
|
|
| Immutable frontend bundles (`cross_dataset.frontend_bundle.v1`) consumed by the calibration |
| diagnostics dashboard's Cross-dataset page: every source (a model × dataset pair, or a standalone |
| dataset) scored against the same Chronicle facts for a jurisdiction. Produced by |
| PolicyEngine/calibration-diagnostics `scripts/run_full_chronicle_evaluation.py` + |
| `publish_cross_dataset_frontend_bundle.py`; every partition carries the run and snapshot ids and |
| a SHA-256 recorded in `manifest.json`. |
|
|
| Layout: `<jurisdiction>/<run_id>/frontend/` and `<jurisdiction>/latest.json` → run id. |
|
|
| Belgium (evaluation-f28ca06a0b0d2baf13c87f2f): sources Microcosm-BE v0.4 × Axiom rules engine, Microcosm-BE v0.4 × EUROMOD |
| BE_2025, EUROMOD BE_2025 on EU-SILC (JRC country report). Microcosm-BE provenance: synthetic |
| Belgian population calibrated to Belgian administrative and national-accounts targets (sums of |
| Chronicle facts; surveys validation-only); support records are a US survey donor pool, reweighted; |
| a Belgian donor pool is the planned upgrade. |
|
|
