Rattibha

Yohan Iddawela

18 Tweets 2 reads Apr 19, 2024

The most efficient way to process nightlights data?
Using Google Earth Engine.
But using it with R can be tough.
So here's a detailed guide on how to do it:
#rstats

By the end of this post, you'll know the steps required to create the following graph of monthly luminosity data in R using Google Earth Engine.
GEE is a fantastic resource—the computations are done on the cloud, so you won't be limited by your computer's hardware.

Step 1:
If you haven't already, you'll need to register for a GEE account.
It can take a couple of days for your account to get approved.

Step 2:
Install the RGEE package using the following instructions from Ricardo Dalagnol's video:
youtu.be

Step 3:
Import your region of interest—i.e. an SF object.
You can download this directly to your workspace using the geodata package.

Step 4:
Import the nightlights data.
In this example, we'll be using monthly nightlights data from the Colorado School of Mines.
You can do this using the ee$ImageCollection() command from RGEE.

Step 5:
Select the relevant band from the nightlights images.
For us, it's the average radiance band (avg_rad).

Step 6:
Filter for the months we're interested in.
It will be January 2014 - December 2022.
We can do this using the filterDate() command from RGEE.

Step 7:
Calculate the zonal statistics for the region of interest.
We can sum up all luminous pixels for the region using the ee_extract() command.
The output will be a wide data frame.

Step 8:
Finally, we can transform the data frame into long format.
Then we can easily the graph the results using ggplot2.
The output is as follows: