Campaign targeting

Two estimates of how many people live in each LGA — the WorldPop grid summed over the LGA boundary, and the census projection the campaigns were planned against — and where they disagree.

WorldPop population

LGA ·

Census projection

WorldPop − census

LGAs apart by more than 5%

WorldPop population per LGA,
WorldPop vs census projection, — % difference (grey: no census projection)

Difference by ,

WorldPop minus the census projection as a share of the projection.

WorldPop against the census projection,

One dot per compared LGA; on the line the two agree. Dots above it are LGAs where WorldPop counts more people than the projection.

By LGA

Both figures are ICRTargetPopulation Groups on the same Location, flattened by the IcrTargetPopulation view (denominator_type = total-population) and paired on location_id and estimate_date. WorldPop is the constrained 100 m grid for the year summed over each LGA boundary by kiln population (a pixel counts for the LGA its centre falls in, so LGAs add up exactly to their state); census projection is the NPC 2006 census and 2022 projection interpolated geometrically to the year, the figure the demo campaigns were planned against, available for the five demo states only. Every other state shows WorldPop alone.

of LGA rows · ▲ more than 5% apart, ▲▲ more than 15%. A large gap is a targeting risk in either direction: a projection above WorldPop means doses and teams planned for people who may not be there; below it means an under-supplied round and coverage that looks better than it is.