Reference

`geo.building_analytics`

domain: geokind: measure

Building-level DERIVED analytics for one geo.building footprint. v1 disaggregates a cell's demographics.population across the buildings inside it, weighted by floor area (area_m2 x levels), so `est_population` sums back to the cell total. A derived MEASURE of a geo.building (x-measures-of), not an observation: `prov.derived_from_refs` pins the exact geo.building + demographics.population inputs it was computed from. est_population is an ESTIMATE (deterministic allocation), never a headcount — see `confidence`. Optional heat/flood fields are reserved for later cell joins; v1 leaves them unset.

Field	Type	Required	Description
`schema`	`—`	✓
`id`	`string`	✓
`building_ref`	`string`	✓	the geo.building this row is derived for (its footprint id)
`h3_id`	`string`		the H3 cell (at the population layer's resolution) whose population was disaggregated
`resolution`	`integer`		H3 resolution of h3_id (the population cell)
`centroid`	`array<number>`		[lng, lat] of the building footprint
`floor_area_m2`	`number`		area_m2 x max(1, levels) — total floor area, the disaggregation weight
`levels_used`	`number`		storeys used for floor area; 1 where building height is unknown (null-height fallback)
`est_population`	`number`		cell population x (this building's floor_area / cell total floor_area). An ESTIMATE, not a headcount.
`heat_lst_c`	`number`		reserved: daytime land-surface temperature at the building's cell (future cell.heat join)
`flood_water_fraction`	`number`		reserved: cell water fraction at the building's cell (future flood join)
`method`	`object`		engine id + inputs + storey height used
`source_ref`	`string`		primary upstream ref.source (inherited from the footprints)
`confidence`	`number`
`prov`	`object`	✓	{ method, derived_from_refs[], observed_at } — pins the exact geo.building + demographics.population versions used
`observed_at`	`string`	✓	ISO8601 time the analytics were computed (valid-time)