Facts request from the microcosm side (microcosm#736 items 9/12; diagnosis: policyengine-uk-data#466, microcosm#738). The consumer contract will add element-level UC declarations once these exist — references then activate automatically on regeneration, and microcosm#729's calibration stage picks them up.
1. UC element-count series (new record sets)
Benefit-unit counts of households whose UC award includes each element:
| Element |
Immediate source |
Notes |
| Childcare |
Published bulletin, no API key: UC statistics to 9 Oct 2025 — 160k (Aug 2025), 170k (Aug 2024) |
Easiest first package; the model runs 2.3× this (uk-data#466) |
| Housing |
Stat-Xplore UC_Households (API key) |
~5.19m model-side, currently unbenchmarked |
| LCWRA / health |
Stat-Xplore UC_Households |
~2.04m model-side |
| Carer |
Stat-Xplore UC_Households |
~0.56m model-side |
Shape: follow the #184 dwp.uc_benefit_units pattern — facts-only, monthly period typing, benefit-unit entity, country geography (K02000001), assertion: observation, provenance to the export/bulletin table. Distinct source concepts per element (or one concept with an element dimension — author's call under the loader guarantees).
2. Monthly history for the family-type and child-count splits (extend existing record sets)
The current feed carries the family-type (single/couple × children) and child-count splits at a single month (2025-05). The caseload series (#184) is monthly, and microcosm's adjudicated calendar-year-average operation needs the same monthly history on the splits to make them sum-consistent with the caseload average (uk-data#466 step 3, uk-data#458; microcosm#736 item 9). Stat-Xplore UC_Households has the series.
Not needed: two-child-limit facts (the April-2025 publication is already packaged; the 2026 abolition retirement is microcosm#736 item 11).
cc @MaxGhenis @vahid-ahmadi — Max's #701 seam comment offered exactly this class of contribution (the #452-thread derivations); the caseload half landed as #184, this is the element half.
Facts request from the microcosm side (microcosm#736 items 9/12; diagnosis: policyengine-uk-data#466, microcosm#738). The consumer contract will add element-level UC declarations once these exist — references then activate automatically on regeneration, and microcosm#729's calibration stage picks them up.
1. UC element-count series (new record sets)
Benefit-unit counts of households whose UC award includes each element:
UC_Households(API key)UC_HouseholdsUC_HouseholdsShape: follow the #184
dwp.uc_benefit_unitspattern — facts-only, monthlyperiodtyping, benefit-unit entity, country geography (K02000001),assertion: observation, provenance to the export/bulletin table. Distinct source concepts per element (or one concept with an element dimension — author's call under the loader guarantees).2. Monthly history for the family-type and child-count splits (extend existing record sets)
The current feed carries the family-type (
single/couple × children) and child-count splits at a single month (2025-05). The caseload series (#184) is monthly, and microcosm's adjudicated calendar-year-average operation needs the same monthly history on the splits to make them sum-consistent with the caseload average (uk-data#466 step 3, uk-data#458; microcosm#736 item 9). Stat-XploreUC_Householdshas the series.Not needed: two-child-limit facts (the April-2025 publication is already packaged; the 2026 abolition retirement is microcosm#736 item 11).
cc @MaxGhenis @vahid-ahmadi — Max's #701 seam comment offered exactly this class of contribution (the #452-thread derivations); the caseload half landed as #184, this is the element half.