Version 1.1.0 of the EconData NIGHTLIGHTS dataset extends our night-time light intensity series beyond South Africa’s borders. Alongside the existing South African series at national, provincial, municipal and ward level, the dataset now includes national-level monthly series for 53 African countries: every African country other than South Africa.
What’s in the data
The series are monthly, running from January 2012 to one month ago. Each value is the average night-time radiance for the area, measured in nW/cm²/sr. The data come from NASA’s Black Marble programme, which derives night-time light intensity from the Day/Night Band of the VIIRS (Visible Infrared Imaging Radiometer Suite) sensor.
Night lights are a widely used proxy for economic activity. They are especially useful where official statistics are sparse, infrequent or published with a long delay, and they give a consistent measure that can be compared across regions and countries.
A change of satellite, and how we correct for it
The satellite behind the series changes partway through. Periods up to and including December 2017 come from Suomi NPP. From January 2018 onward, the data come from NOAA-20, which carries the same VIIRS instrument family on a newer platform. NASA’s delivery of Suomi NPP data ends in November 2026.
The two satellites’ raw readings differ systematically, in ways that have nothing to do with actual light levels on the ground. To keep the full history from 2012 on a comparable scale, a correction factor is applied to every value from 2018 onward. Values before 2018 are the as-reported readings and are not corrected.
The correction differs by geography:
- South Africa, which is treated as a single climate, uses a correction that varies by calendar month.
- The rest of Africa uses one annual correction for each of five broad climate zones. Month-to-month variation in weather and biomass burning would otherwise swamp the calibration signal in several zones.
Corrections for the most recent one to two years may be refined further as NASA’s multi-year reprocessing of the underlying imagery matures, so recent values may be revised in later releases.
Getting the data
The dataset is available on EconData as dataset NIGHTLIGHTS. In R, using the econdatar package:
nightlights <- econdatar::read_dataset("NIGHTLIGHTS")
For details on the concepts, see the spatial data user guide. Below is a chart of Ghana’s national series.
