This section gives you additional examples to explore after the guided burnt forest area case study. The notebooks show common access patterns for EarthCODE-hosted research data products, especially Zarr and Parquet assets opened directly from object storage.
Use these notebooks as patterns: identify the dataset URL or catalog item, open it with the appropriate Python library, inspect dimensions and variables, then subset or visualise the data for your research question.
Recommended During the Workshop¶
SEASFIRE data cube Opens a fire-focused Zarr data cube with Xarray, uses Dask-backed arrays, subsets by area and time, and connects the catalog discovery workflow to the burnt-area case study.
Access Antarctica Data Opens multiple Antarctica analysis-ready Zarr cubes, merges them into a single Xarray view, subsets an ice shelf region, and visualises basal melt and calving-front variables.
More Direct Zarr Examples¶
SEN4Amazonas Opens a tree-cover-change Zarr cube, selects a small Amazon region, computes the subset, and visualises time-varying forest change.
SMART-CH4 Opens a methane isotope signal Zarr dataset and plots selected variables for quick inspection.
WAPOSAL Opens an altimetry-style Zarr product, inspects trajectory metadata, and plots a selected track on a map.
ARCEME Opens an ARCEME Zarr data cube directly from object storage and inspects its dataset structure.
AGRICEM Opens an agriculture-focused Zarr data cube and plots a selected band/time slice.
Parquet and Tabular-Geospatial Examples¶
Lightning2EarthCARE data access Demonstrates Parquet access with GeoPandas and DuckDB, including spatial and column filtering.
Lightning2EarthCARE visualisation Visualises lightning observations and EarthCARE-related products. Some authenticated EarthCARE product cells depend on a local helper module and may need instructor setup before running end-to-end.
What to Compare¶
Across the examples, look for these recurring steps:
Where the data URL or catalog entry comes from.
Whether the data opens with
xarray,geopandas, orduckdb.Which operations stay lazy and which trigger computation.
How small spatial or temporal subsets make large remote datasets practical to explore.