Use this code to reproduce graph.

• Analysis-ready dataset of monthly nighttime lights derived from the VIIRS instrument aboard the SNPP satellite.

• Nightlight is calculated at the administrative unit level (ADM 0~2) for 208 countries and territories, totaling 52,702 administrative units worldwide.

• Administrative units are each represented by a time series (sequence of monthly observations) for the period between January 2012 and May 2024.

• The dataset is updated regularly and hosted publicly on AWS.

Click here to access the dataset.

Individual .csv files, one per administrative unit, adhere to the following schema:

¹See geoBoundaries for details.

1. Monthly composites (DNB_npp_*-slcorr_v10_*.avg_rade9.tif) and cloud-free masks (DNB_npp_*-slcorr_v10_*.n_cf.tif) are sourced from Light Every Night (LEN) on AWS.
2. nodata pixels (-1.5 and -999.3) are set to numpy.nan.

1. Global boundaries with three levels of administrative divisions (ADM0~2) are sourced from geoBoundaries.
2. Level 2 boundaries are mapped to their parent level 1 polygons via intersection.
3. This vector layer is used below to extract zonal statistics from the raster layer.

1. Zonal stats (min, max, etc.) at the administrative unit level are extracted from the raster layer above using rasterstats library with the all_touched flag set to False.
2. cloudFree column is calculated from the cloud-free mask using the same approach described in step 1.
3. Extracted time series data are written to individual .csv files and uploaded to S3 on AWS.

1. Orphan ADM2 boundaries listed under Null due to polygon intersection failure. See here for an example.

2. Missing data: data missing from the source (03/2014, 08/2021, 06/2022 and 08/2022 are replaced with numpy.nan in the final dataset.

3. Missing masks: cloud-free masks missing from the source (10/2017 and 11/2017) are replaced with numpy.nan accordingly.

This dataset is made available under the CC BY 4.0 license, requiring acknowledgment in any products that use it.

Written works:

Najjar, A. (2024). Global scale nightlight time series dataset. https://nightlight.eoatlas.org.

Online use:

Make sure to mention Earth Observation Atlas as the source and link to https://eoatlas.org.