These endpoints retain a historical reviewed estimate alongside its actual current engine state, evaluated 2026-10-27. Preserve both. An expired or replaced observation remains inspectable; its review flag does not keep it admitted. Query parameters do not filter these static files.
GET /api/v1/heat-exposure.json
A year of heat, with its monthly detail.
63 destination bindings across 62 physical estimated areas; 63 exact historical annual inputs currently admitted. Evidence reviewed 2026-10-26.
Keep weather 2025, population weights 2020, boundary 2015, UTC days and the 38°C outdoor UTCI threshold. The unit is person-days/person. The twelve monthly contributions are dependent parts of one annual input, not twelve separately admitted metrics.
The provider revision is not returned: preserve null and the exact original chart-response archive hash. This is regional model weather, not a forecast, climate normal or clinical risk estimate. Original C3S/ECMWF chart facts retain their data licence, modification attribution and responsibility notice; no provider frontend-code grant is implied.
GET /api/v1/rental-vacancy.json
One Canadian survey cohort.
4 destination bindings in October 2025; 4 exact historical survey inputs currently admitted. Evidence reviewed 2026-10-26. This subset does not establish worldwide vacancy coverage.
Preserve the combined private apartment and row universe, official CMA 2021 identifiers, source quality flags and one-decimal published precision. Calendar bounds identify the survey month; no exact fieldwork day or independently observed vacant-unit count is invented.
The engine ages month-precision evidence conservatively from its first possible day. The API keeps the original temporal basis and published-rate derivation. CMHC conditional data terms and acknowledgements accompany each record; this API grants no additional reuse rights.
GET /api/v1/housing-supply.json
A construction flow, with its denominator.
Calendar 2025 gross new dwelling completions per 1,000 preliminary July 1 residents. 4 of 4 exact reviewed Canadian inputs currently admitted; evidence reviewed 2026-10-26.
CMHC's structural milestone may allow up to 10% unfinished work in some circumstances; it is not occupancy certification. Preserve all intended markets, original integer operands and the preliminary population status.
Native publication notes assign CMA2021 scope. Keep the completion CSV's 2011-prefixed DGUID, the population's national CMA2021 DGUID and the boundary's provincial-part DGUID in their original roles. CMHC conditional terms and separate Statistics Canada rights follow the data; this API grants no additional reuse rights.
GET /api/v1/life-expectancy.json
The original cell, with its qualifications.
ABS Persons life expectancy at birth for 2022–2024, original one-decimal cells on native GCCSA2021 functional regions. 4 of 4 exact historical inputs currently admitted; evidence reviewed 2026-10-26.
The derivation keeps original workbook/header/cell references, native EPSG:7844 geometry and the reviewed destination-point adapter. GCCSA algorithm, terminal age, overseas-death rule and source CI/SE remain null; the SA4 method and a male/female average are not substituted.
Preserve CC BY 4.0 source credit, adaptation responsibility and exact current identity, including reviewed_at. A historical review does not override expiry or replacement, or establish quarterly comparison clearance.
GET /api/v1/road-deaths.json
An event count with a matching region.
7 of 7 original Calendar 2024 · crash year inputs currently admitted. Preserve motor-vehicle involvement, territorial crash-year allocation, source 30-day classification, provisional ARF, exact fatal-person counts and the compatible Vintage 2024 midyear population denominator.
The source covers complete county-based MSAs, rather than a city centre or the old PM footprint. Unknown injury categories and individual death times remain unknown. Missing point coordinates do not remove county-qualified cases. Later-vintage comparisons are separate; no resident or visitor injury probability is inferred.
Large original archives retain ordered binary parts and a whole-file manifest. A part is not a standalone ZIP. Exact reconstruction, source-specific government-data terms and attributions, reviewed method versions and current/historical status accompany the public records.