Google Earth Engine (GEE) brings together large collections of satellite imagery and Earth observation data with cloud-based tools for analysis and visualization. In this session, we’ll explore how GEE can be used to move from large-scale satellite data to actionable insights, from working with hyperspectral imagery and time series to applying machine learning and new satellite embedding datasets for geospatial analysis.

------------------------------------------------------

Topic: From Pixels to Intelligence: A Hands-On Guide to Google’s New Satellite Embedding Dataset

Speaker: Guneet Mutreja, Doctoral Researcher at German Aerospace Center

Description: Satellite imagery offers a firehose of data, but how do we move beyond simple pixels to genuine understanding? The answer lies in embeddings—a concept developers know well from the world of NLP. In this demo-driven session, Guneet will introduce Google's new Satellite Embedding Dataset, which functions like "Word2Vec for our planet." We'll explore how these pre-computed embeddings, powered by foundation models like AlphaEarth, can enable new approaches to complex geospatial analysis in Google Earth Engine. Specifically,

- Perform semantic searches (e.g., "find all solar farms like this one") with a single click.

- Build accurate land-cover maps in seconds using few-shot classification.

- Leverage vector similarity to unlock new analytical capabilities, all without training a complex model from scratch.

------------------------------------------------------

Topic: Leveraging Hyperspectral Data in Google Earth Engine

Speaker: Dr. Emil Cherrington, Research Professor at University of Notre Dame Environmental Research Center

Description:

Related to ongoing Earth observation research at the  University of Notre Dame Environmental Research Center (UNDERC), in his 30 minute session, Emil will highlight the utility of hyperspectral data available through Google Earth Engine. He will specifically explore a GEE-based Toolkit he originally developed for data from NASA's PACE (Plankton, Aerosol, Cloud, ocean Evosystem) hyperspectral mission, in the context of the NASA EarthRISE initiative. Where PACE provides ~daily hyperspectral coverage over the entire globe, the session will cover time series analysis and classical machine learning approaches for working with big Earth data from PACE's Ocean Color Instrument (OCI).

Resource: PACE_OCI_Toolkit https://github.com/BzGEO/PACE_OCI_Toolkit

Cherrington, E. (2026). PACE OCI data Toolkit (Version 1.0.0.6) [Computer software]. https://doi.org/10.5281/zenodo.18382540