Recently I saw a post about the relatively new NASA Black Marble (night time satellite imagery) datasets being released. Shortly after a few projects popped up such as Digital Geography’s slider comparison and ESRI’s Difference Detection map. ESRI also has a story map and blog post with more detail regarding their process. There were a couple of quirks to me and I decided to check out a few of these setups myself and decided on a hybrid I thought was most informative.
I preferred the Leaflet setup of Digital Geography (although OpenSeaDragon is amazing and supports basic map visualization). Leaflet opens up more analytical capability that couples with maps and is much easier on lighter connections compared to a standard OpenSeaDragon setup. I also personally prefer a switcher to a slider as my personal preference for comparison.
However ESRI’s project does have an excellent layer to it.. the “difference” highlights. Although flipping or sliding allows you to compare visually, this method can help highlight important changes very noticeably while exploring and has the added bonus of producing rough estimates with raster analysis. Theoretically you could calculate the number of pixels with light changes to determine a rough area of added light versus an area of receding light. I, personally, did have some reservations to the results that ESRI posted. As an example, a closeup with Puerto Rico showed that it appeared quite magenta to indicate receding light. However, if you compare the lighting on the imagery sets there is definitely some receding… but it doesn’t seem quite so pronounced as the graphic indicated to me. Due to my misunderstanding of the intensity I was expecting a much darker island which is definitely not the case. I predict this was caused by Arc’s inability to handle a wide variety of transparency issues. ArcGIS Pro has a lot of improvements and should have been able to handle this, but there isn’t necessarily a scale for such on the map as I see it.
As such I wanted to have the same highlighting, but with a scale based on transparency that would highlight changes based on their intensity.. all wrapped in a clean Leaflet setup. Below is my end result. The light differences aren’t as vibrant for a “wow” factor, but should show the actual changes in light intensity more accurately. Places such as North India and Syria still have more noticeable patches. Perhaps I will update it to range from vibrant orange to red and vibrant purple to green to show intensity without taking away any visual appeal. I’d have to play around..
NASA – Night Light Comparison is a self-hosted* project I put together that has these factors. A small interactive version is also below for you to check out. I’m currently collecting my material and exploring my workflows to explain my process from data download to a completed web map. It will all be in my next post, so I hope you check it out as well!
* The label layer in this project is provided by an ESRI tile service.
The Dark Basemap is provided by CartoDB.