Open Satellite Data for Disasters and Investigative Journalism

Maxar Tech's DigitalGlobe is part of the disaster charter. 

The recent ammonium nitrate explosion in Beirut, bushfires in Australia in early January and the ongoing fight against Covid-19 have all but one thing in common; the need for facts, data and analysis to understand and act responsibly. One of the better ways to mine data is through satellites and while sub meter resolution data might cost thousands of dollars, there are other ways to retrieve meter-level information for free. Today, I will be looking at some of the resources online that brings space right down to your fingertips. 

One thing to note is that several people have already compiled such lists. GISGeography has list of 15 free satellite imagery data sources that interested parties can explore. It extends on the article I wrote in 2017 about satellite data for Nepal which also has information on different indices that can be done with band math to classify images. However, with the advancement of Machine Learning (ML), specifically on Deep Learning, we will probably see a lot of band math being used along ML for image classification and segmentation purposes. 

GISGeography enlists USGS Earth Explorer, which has data from its long standing LandSat program. I had used it to write my paper back in 2015 after the earthquake in Nepal but the interface is a bit old for 2020. Most interfaces now allow you to process image on the website itself without needing to have an expensive, heavy software like ESRI. Everything is done on the cloud and all you need is to do downlink the results. 

Case in point Google's Earth Engine which has datasets from LandSat and the popular Sentinel programs. You will need a google account and some Javascript to code but you will find plenty of tutorials on YouTube to get along with your first project just fine. Here's an excellent article on studying the river path using the Earth Engine platform on medium. A similar powerful tool I have come across that's free (although if you want, you can purchase high resolution data on the spot) is EOS' landviewer application. You can do all sorts of things without leaving the website and is a powerful tool if you want immediate answers to your satellite data questions. 

Sentinel Open Access Hub is dedicated to Sentinel data with an interface similar to USGS Earth Explorer. I have noticed Sentinel's EO Browser to be much more powerful. However, you will need to sign in and there's a time limit. A higher resolution data interface is provided by Planet where you can create timescale data such as Nepal's Sisdol landfil site to understand how it's making an impact on the environment. You will receive free two week access to the data. 

An excellent article to start understanding how satellite resolution matters, how to program and use it for investigative journalism is written by Techjournalist on medium. The article has a top to bottom guide on forest fire detection, tracking terrorists and finding changes on a water body. All you have to do is learn from the examples provided (Python on Jupyter notebook as I remember) and then apply to your own problems. 

It's also worth noting that through the UN SPIDER/UNOSAT program, space agencies who have signed on to the program have an obligation to provide free data. Any national or international organization in Nepal can request by call and form submission to UNOSAT. If the emergency is valid, UNOSAT will then trigger the International Charter Space and Disaster. Once the Charter is triggered, the data will then flow back to the organization that requested it. Companies like Planet and Digital Globe are also part of the charter so you can expect higher resolution data support. 

You can watch my full talk on using Satellites for Disaster Prevention and Response below: (edit: looks like you will have to go to YouTube to watch it as third party website viewing is disabled)


  1. Thanks for sharing all the valuable resources to dig up the satellite Data. Also, the video embedding has been enabled now.


Post a Comment

Popular Posts