Context: we analyze economic changes in Colombian municipalities between 1993 and 2020.

In this post, we focus on how to download our “raw data”, i.e. the nighttime light data for Colombia.

Step 1: Download nighttime light series from NOAA National Centers for Environmental Information (NCEI). Particularly Version 4 DMSP-OLS Nighttime Lights Time Series (DMSP) – The DMSP annual composite data contain average radiance values of cloud-free coverages, reflecting the persistent lights from cities, villages, and roads, with a spatial resolution of about 900m, and a temporal coverage of 1992 to 2013 – and VIIRS data which is available from 2013 on and with is a finer spatial resolution of 450m approximately. We will later do a post on the particularities of VIIRS, here we focus on DMSP.

Once we are in NOAA website, we will see two different tables. We use “Average Visible, Stable Lights, & Cloud Free Coverages” table. Why? because clouds is a major constraint in night light analysis and this table contains cloud free images.

We click on every year and wait until the file is downloaded. We save all the files in a folder “Nightlight Analysis” and within that folder, we create one new sub-folder for each of the years.

Each of the year contains eight files compressed in a format .tar

Once we place each of the .tar files in their respective year folder, we extract the file and obtain the following files. The most important for our analysis is F162008.v4b_web.stable_lights.avg_vis.tif At the end of this exercise you should have as many avg_vis.tif as years you want to analyse. Then you will be ready to perform the next step: uploading nighttime light file to a GIS software, and convert them from raster to points before we build our dataset.