Sample Spectral Clustering for Land Use Classification Using Sentinel-2 L2 Data

Recording

About the webinar

This session highlights outcomes from the EOPF Hackathon held during BiDS25, where participants explored innovative approaches to processing Sentinel data in Zarr format.

One of the hackathon teams presents their project, “Sample Spectral Clustering for Land Use Classification Using Sentinel-2 L2 Data”. The team walks through their workflow, shares technical methods, and discusses key insights gained while experimenting with the EOPF Sentinel Zarr datasets.

What to expect

Through a live notebook demonstration, participants learn how to:

• Apply spectral clustering techniques for land use classification.
• Work with Sentinel-2 Level-2 data in EOPF Zarr format.
• Understand the differences and advantages of cloud-native EO data formats.
• Gain insights from real-world hackathon experiments and lessons learned.

Background

The webinar contributes to the broader effort to promote adoption of the EOPF Zarr data products and the EOPF Sample Service. The goal is to support a growing community of early adopters experimenting with new, cloud-ready Earth observation data formats, and to show how these developments enable faster, scalable and more collaborative EO data processing.

Who should watch

This session is tailored for intermediate to advanced Earth observation users, especially those interested in Sentinel-2 data, machine learning for land classification, or cloud-native EO processing.

Programme