WorldPop — Available datasets

earthlens curates the WorldPop population & demographic product families. Each product (a top-level REST alias) offers one or more sub-aliases, selected by the (constrained, unadjusted, resolution, scope, generation, level) tuple. The default selectors resolve to the classic unconstrained 100 m per-country series (pop → wpgp).

This page is generated from the bundled worldpop_data_catalog.yaml; regenerate the upstream index and re-validate it with the catalog tool.

Curated product / sub-alias matrix

product	friendly	kind	demo	unit	sub-alias	constr	unadj	res	scope	gen	level	years
age_structures	age_sex, age_structure, demographics	mixed	True	people/pixel	aswpgp	False	True	100m	countries	R2021	national	2000-2020
					aswpgponekm	False	True	1km	global	R2021	national	2000-2020
					ascic_2020	True	True	100m	countries	R2021	national	2020
					ascicua_2020	True	False	100m	countries	R2021	national	2020
births	births_count	raster	False	births/pixel	bic	False	True	100m	countries	R2021	national	2000-2020
dependency_ratios	dependency_ratio, dependency	raster	False	ratio	drwc	False	True	1km	global	R2021	national	2000-2020
dug	degree_of_urbanisation, degurba	raster	False	class	dug_g2_v1	True	True	100m	countries	R2025A	national	2015-2030
future_pop	projections, population_projection, ssp	raster	False	people/pixel	FPP_v02	False	True	1km	global	R2025A	national	2025-2100
gbsg	built_settlement_growth, settlement_growth	raster	False	class	bsgm	False	True	100m	countries	R2021	national	2000-2020
pop	population, population_counts, ppp	raster	False	people/pixel	wpgp	False	True	100m	countries	R2021	national	2000-2020
					wpgpunadj	False	False	100m	countries	R2021	national	2000-2020
					wpic1km	False	True	1km	countries	R2021	national	2000-2020
					wpicuadj1km	False	False	1km	countries	R2021	national	2000-2020
					wpgp1km	False	True	1km	global	R2021	national	2000-2020
					cic2020_100m	True	True	100m	countries	R2021	national	2020
					cic2020_UNadj_100m	True	False	100m	countries	R2021	national	2020
					G2_CN_POP_R25A_100m	True	True	100m	countries	R2025A	national	2015-2030
					G2_CN_POP_R25A_1km	True	True	1km	countries	R2025A	national	2015-2030
					G2_MOS_POP_R25A_1km	True	True	1km	global	R2025A	national	2015-2030
pop_density	population_density, density, pd	raster	False	people/km2	pd_ic_1km	False	True	1km	countries	R2021	national	2000-2020
					pd_ic_1km_unadj	False	False	1km	countries	R2021	national	2000-2020
pregnancies	pregnancies_count	raster	False	pregnancies/pixel	individual_countries	False	True	100m	countries	R2021	national	2000-2020
pwd	population_weighted_density	raster	False	people/km2	pwd_national_1km	False	True	1km	countries	R2021	national	2000-2020
					pwd_national_100m	False	True	100m	countries	R2021	national	2000-2020
					pwd_subnational_1km	False	True	1km	countries	R2021	subnational	2000-2020
					pwd_subnational_100m	False	True	100m	countries	R2021	subnational	2000-2020
urban_change	urban, urbanisation	raster	False	class	ucic	False	True	100m	countries	R2021	national	2000-2020

demo = True marks a demographic product whose per-cohort rasters are also tabularised (only age_structures). kind = mixed follows.

Age/sex cohorts (age_structures)

Each age_structures year ships 36 GeoTIFFs = 2 sexes (m, f) × 18 age bands. The band lower bounds are:

0 (<1), 1 (1–4), 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80 (80+)

The filename pattern is {iso3}_{sex}_{age_low}_{year}.tif (e.g. ken_f_0_2020.tif); the AOI population sum per cohort is written to a tidy table — see Usage → Demographic tables.

Global mosaics & archive products

The scope="global" mosaics of pop / age_structures (wpgp1km, aswpgponekm, G2_MOS_POP_R25A_1km, …) are fetchable: the backend resolves the per-year whole-world GeoTIFF via the hub's ?id= detail endpoint, downloads it, and crops to the AOI. WorldPop has no server-side subsetting, so each global 1 km mosaic is a ~1.1 GB download per year.

dependency_ratios (drwc, archive: "7z") and future_pop (FPP_v02, archive: "zip") are distributed as multi-file archives, not per-year GeoTIFFs; the backend downloads the archive, extracts its GeoTIFFs (py7zr for .7z), and crops them to the AOI. dependency_ratios is small and per-continent (Asia / Africa only, year 2010, three ratio variants — resolved from the AOI's continent). future_pop is per-SSP and ~4 GB per scenario, so it needs allow_large_archive=True + ssp= to opt in. Both require the [worldpop] extra.

Covariates (54 layers)

The hub's covariates family — VIIRS/DMSP nightlights, SRTM slope/elevation, distance-to-roads/water, ESA-CCI land-cover distances, Global-2 built-up surface/volume, building patterns, … — is fully curated: each of the 54 layers is its own product (selected by its id, e.g. variables=["cviirs"]), routed through the shared covariates REST endpoint (rest_alias: covariates). They are per-country GeoTIFFs and use the normal country fetch path. List them with Catalog().describe(<id>) or the audit tool; each product carries the hub's title as its description.

EarthLens(data_source="worldpop", variables=["cviirs"],  # VIIRS nightlights
          start="2012", end="2016", fmt="%Y", aoi="KEN",
          lat_lim=[-4.7, 5.0], lon_lim=[33.9, 41.9], path="out/").download()

Catalog tooling

tools/worldpop/ mirrors the GEE / CHC / tropycal catalog tooling — a refresh, an audit, and a probe:

# refresh: crawl the hub and write the available_products index (every alias + sub-aliases)
python tools/worldpop/refresh_worldpop_catalog.py refresh -o available_products.yaml

# refresh: structural check of the curated catalog (offline) + upstream existence (--live)
python tools/worldpop/refresh_worldpop_catalog.py validate --live --check-files

# audit: diff the curated catalog vs the live hub; --strict exits non-zero on drift
python tools/worldpop/audit_worldpop_catalog.py --strict

# probe: capture the live REST shape (aliases, sub-aliases, a sample record) to JSON
python tools/worldpop/probe_worldpop_rest.py --alias pop --subalias wpgp --iso3 KEN -o probe.json

The audit report flags curated products / sub-aliases the hub no longer serves and new upstream products that aren't curated (excluding the deliberately out-of-scope specialty families). The Catalog also exposes a runtime health() hygiene report and a describe(product) introspection helper, matching the other backends' catalogs.

Licence

All WorldPop data is CC-BY-4.0 (attribution). The required citation is carried per dataset in the REST record (citation field); the canonical licence is at https://hub.worldpop.org/data/licence.txt.