Descartes Labs

Sign up

Please sign up with your university email address at iam.descarteslabs.com. You need to log in once to accept the terms and conditions. The tools available then are extremely limited by default. Once you signed up, please email the office that you have done so. We forward the emails altogether to Descartes Labs who will then give us as a group access to their APIs, non-programmatic interfaces, their catalog products with several satellite imagery sources, and the hosted Workbench development environment.

Getting started:

• Take a look at our onboarding video series
• Walk through the articles in Knowledge Base
• Watch this webinar that provides an overview for Workbench, the hosted development environment
• The guides and examples in Workbench under the ~/example_notebooks folder can be especially helpful, as you can walk through the code at the same time as you learn the tools.
• There is a user forum at GIS Stack Exchange (just tag "descartes-labs-platform" in any question you ask and they'll be notified). This is useful for platform-specific questions. Check to see if your question has already been asked/answered on the GIS Stack Exchange. If not, you can submit and tag your question there and someone on our team (or within the GIS Stack Exchange community) will respond.
• For other questions, customer support can be contacted via email at [email protected] or you can submit a support ticket through support.descarteslabs.com

Landing Page

Your IAM homepage is a helpful page to bookmark as it contains links to other areas and functionalities of the platform. On the left-hand bar you will see:

• Workbench: hosted development environment with pre-loaded example scripts for programmatic users
• Catalog: raster imagery products available on the platform, as well as any you have uploaded yourself or that have been shared with you
• Viewer: tool to easily visualize Catalog raster products
• Task Monitor: monitor ongoing computing jobs submitted to the Tasks API
• Docs: API documentation, best practices, and additional example code

Setting up Workbench

• Log in to your account here: IAM
• Navigate to Workbench, the hosted JupyterLab environment, by selecting "Launch Workbench":
  • Start a server. The topmost "Model Exploration/Development" option is going to be the best for now. Other options using older versions of the Python client are also available, as well as a GPU-enabled "Model Training" counterpart for each of these, but you'll likely use the first option most of the time.
  • Run through the 01_initial_setup.ipynb notebook located in the ~/example_notebooks folder. This will ask you to log in once again and present a token. Click the "Copy Token" button and paste in the text field within the notebook. Press return and your email or name should be printed out. You should only need to do this once, but may need to rerun this if at any point you are getting an error similar to could not find client_id
• Test your permissions by running through the 02_getting_started.ipynb notebook within the same ~/example_notebooks folder.
• Try out the examples in the ~/example_notebooks/examples/ folder to get an idea of some of the many analysis examples you can implement on the DL platform.
• The ~/example_notebooks/guides/ folder takes you through some of the DL APIs in closer detail.
• NOTE: The ~/example_notebooks folder should be READ ONLY. Any work saved there will not persist when you restart your server. For this reason, please save any custom code outside of that folder. We suggest creating a folder for your work within the root directory.

Documentation

The docs are broken up into several sections, in increasing level of detail:

• Tutorials: Tutorials are end-to-end solutions built on their software. These are adapted from their Applied Scientists’ research projects and are a start if you want to learn how their team uses DL software.
• Examples: Examples demonstrate smaller pieces of functionality and are geared toward composability. Let them know if there’s a topic you would like to see.
• Guides: Guides contain best practices and more context than Tutorials or Examples. If you’re wondering “How should I use Tasks?”, you should check out the Tasks Guide.
• API Reference: Lastly, the API reference is there when you need to dig into usage subtleties.