Your Earth, OnDemand
Earth OnDemand’s newest feature allows users to discover imagery and export their query to EarthAI in an analysis-ready function.
Eight months ago, Astraea released the first version of Earth OnDemand with the hopes of democratizing access to the Earth’s greatest data set — the Earth itself.
Last week, we released a new version with one of the most important features yet. Before we get to that, however, it's worth taking a moment to discuss why we built Earth OnDemand in the first place.
Three years ago, Astraea’s founders set out to build information products from publicly available satellite data. As they embarked on this mission, they quickly realized that this was going to be hard. Too hard in fact. To be clear, they were more than capable of the task. The team included computer scientists, data scientists, and a remote sensing guru.
They encountered several significant challenges:
- Acquiring the data required going to numerous different web portals.
- Downloading the data took a long time.
- After they downloaded the data, they found that many of their images were covered in clouds.
- Once they had gathered a set of imagery without clouds in the key locations, they had amassed gigabytes of data — choking up their laptops and GIS tools.
- After they brushed up on AWS and stood up a cloud compute platform for processing all this data, they realized the tools for doing image processing at scale were insufficient.
- Commercial data were expensive and sometimes just as difficult to work with.
It was just one problem after another. To solve any significant question using satellite imagery data could easily cost hundreds of thousands to a million dollars.
While there are a few million-dollar problems out there, there are many more ten thousand dollar questions.
But how could people solve $10,000 questions if it cost $500,000 to answer them? Thus the idea of the EarthAI platform was born.
The Astraea team decided to build a platform to address all of these challenges and thereby enable a wide range of users to take advantage of Earth-observation data.
Over the next few years, this idea has evolved, but the team has remained true to this vision. Astraea has built the EarthAI platform to address these challenges. Earth OnDemand was the first of these tools to be launched in July 2019.
The goal of Earth OnDemand was to make it easy to discover Earth-observation data, conduct visual analysis on it, and export it to other tools for further analysis.
We knew many people wanted to answer questions like:
- What areas did the images cover?
- When were they taken?
- How much cloud cover is there?
- What did the image look like?
- What has changed over a period of time?
Spatial analysts wanted to download these data for analysis in GIS tools and data scientists wanted to build analytics on analysis-ready data sets in notebooks.
Since July, the team has worked to build a set of features to address these needs. This brings us to the present where we have satisfied one of the last of these requirements — the ability to export query results from Earth OnDemand directly into EarthAI Notebook for analysis.
What does query export mean for Earth OnDemand users?
It means that now analysts can go end-to-end with their analysis using the EarthAI platform. They can discover data in Earth OnDemand, conduct preliminary visual analysis of these images, see how far back and how many images satisfy their criteria, quickly compare before and after views of the data, and then take the search results and load them as analysis-ready data in EarthAI.
From there, the sky is the limit for their analysis. They have a JupyterLab Notebook full of all the usual spatial goodies ready to analyze their data.
Other parts of our platform include our open-source RasterFrames library for analyzing imagery at scale and EarthAI Notebook, a hosted JupyterLab environment for analyzing imagery data.
RasterFrames® brings together Earth-observation (EO) data access, cloud computing, and DataFrame-based data science…
To learn more or start a free trial, go to our EarthAI Platform page: