WorldPop — covariates: VIIRS night-time lights (live)¶

The 54 WorldPop covariate layers (nightlights, slope, elevation, distances, built-up, …) are each a selectable product. Here we pull the VIIRS night-time lights (cviirs) for Kenya and crop to a bbox.

Setup¶

Consolidate the imports up front. earthlens.worldpop provides the covariate Catalog; EarthLens is the unified download entry point; pyramids reads the written rasters back; tempfile gives us a scratch output directory.

In [ ]:

import tempfile

from pyramids.dataset import Dataset

from earthlens.earthlens import EarthLens
from earthlens.worldpop import Catalog

import tempfile from pyramids.dataset import Dataset from earthlens.earthlens import EarthLens from earthlens.worldpop import Catalog

1 · Find the covariate layers¶

Load the WorldPop catalog and filter its products down to the covariates REST family. Counting them confirms the full covariate set is available before we request a specific layer.

In [ ]:

cat = Catalog()
covariates = [
    p for p in cat.available_products() if cat.get(p).rest_alias == 'covariates'
]
print('covariate layers:', len(covariates))

cat = Catalog() covariates = [ p for p in cat.available_products() if cat.get(p).rest_alias == 'covariates' ] print('covariate layers:', len(covariates))

Describe the nightlights product¶

describe() returns the metadata for a single product; here we read the human-readable description of the VIIRS night-time lights layer (cviirs) we are about to download.

In [ ]:

print('cviirs ->', cat.describe('cviirs')['description'])

print('cviirs ->', cat.describe('cviirs')['description'])

2 · Download VIIRS nightlights for Kenya¶

WorldPop writes files to disk, so we point the download at a temporary directory. We request the cviirs covariate for Kenya (KEN) across 2012–2016 and crop to a small bbox around the equator.

In [ ]:

d = tempfile.mkdtemp()

d = tempfile.mkdtemp()

Build the request¶

Construct the EarthLens request on its own line so every parameter — the country AOI, the year window, the %Y date format, and the lat/lon crop limits — is easy to read and adjust.

In [ ]:

covariates_request = EarthLens(
    data_source='worldpop',
    variables=['cviirs'],
    aoi='KEN',
    start='2012',
    end='2016',
    fmt='%Y',
    lat_lim=[-1.0, 1.0],
    lon_lim=[36.0, 38.0],
    path=d,
)

covariates_request = EarthLens( data_source='worldpop', variables=['cviirs'], aoi='KEN', start='2012', end='2016', fmt='%Y', lat_lim=[-1.0, 1.0], lon_lim=[36.0, 38.0], path=d, )

Run the download¶

download() mosaics and crops the source tiles per year and returns the list of written file paths. We print their names to confirm one raster per available VIIRS year.

In [ ]:

paths = covariates_request.download(progress_bar=False)
print('written:', sorted(p.name for p in paths))

paths = covariates_request.download(progress_bar=False) print('written:', sorted(p.name for p in paths))

3 · Open a result with pyramids¶

One cropped raster was written per available VIIRS year (2012–2016). We open the earliest one and inspect its CRS and shape to verify the crop.

In [ ]:

first_path = sorted(paths)[0]
ds = Dataset.read_file(str(first_path))
print('EPSG:', ds.epsg, 'shape:', ds.shape)

first_path = sorted(paths)[0] ds = Dataset.read_file(str(first_path)) print('EPSG:', ds.epsg, 'shape:', ds.shape)