Which districts face the most overlapping hazards (flood, coastal, heat, air), and what is each district's exposure on each?

Multihazard exposure. Part of what the twin can answer. Every figure below is the value the route serves, with its source; the raw response is at /api/twin/brain/multihazard.

Descriptive, sourced: no model

Basis: sourced records (URL, sha256 or locator per row), no model.

Each hazard is converted to a within-country percentile rank (0 to 100, 100 = most exposed district); exposure_rank is the mean of the four percentile ranks with equal weights. Ranks are relative across Bangladesh's 64 districts, not absolute risk scores. Method: scripts/gee/multihazard_districts.py.

Most exposed district, Dhaka: 97
People in the 100-year floodplain: 82,403,842
People breathing PM2.5 above 50: 131,076,105
Cropland in the floodplain: 4,656,769.699999999 ha

The most exposed district on each hazard

Per-hazard leaders

Hazard	District	Value
Flood	District: Dhaka	Value: 6,188,271
Coastal	District: Dhaka	Value:

Which hazard dominates where

Dominant hazard is the hazard with the highest percentile rank; ties break in the fixed order flood, coastal, heat, air.

Districts where flooding dominates: 20
Districts where the coast dominates: 12
Districts where heat dominates: 18
Districts where dirty air dominates: 14

Twelve most exposed districts

Top 12 by exposure rank, with each hazard's held value

District	Exposure rank	Dominant hazard	In the floodplain	Below 5 m	Daytime heat, C	PM2.5
Dhaka	Exposure rank: 97	Dominant hazard: flood

All 64 districts

Districts: exposure rank, dominant hazard and each hazard's held value

District	Exposure rank	Dominant hazard	In the floodplain	Below 5 m	Daytime heat, C	PM2.5	Cropland in the floodplain, ha
Dhaka	Exposure rank: 97

Where these figures come from

The source collections

Source	Publisher	Unit	Caveat
Multihazard district exposure join (64 districts)	Publisher: BPL GEE reductions, joined by scripts/gee/`multihazard_districts.py`	Unit: one row per district (ADM2_NAME)	Caveat: The four reduction files were computed server-side in Google Earth Engine (getInfo); no primary raster is saved on this host, so no row names its image or date. The table carries no observation date; the sha256 pin is the vintage.

Not held

Validated exposure, e-GP money at risk and cyclone outputs: They belong to the hazard, climate and cyclone brains; this route links to them and does not repeat them.
Raster provenance (image, date, scene): The reductions were computed server-side; no primary raster is saved on this host, so no row can name its image or date.
Observation dates: Neither the join nor the reductions store one; the sha256 pin is the vintage.

Caveats

The four reduction files were computed server-side in Google Earth Engine (getInfo); no primary raster is saved on this host, so no row names its image or date. The table carries no observation date; the sha256 pin is the vintage.