datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
hydrolakes
Quick Links
Website
GitHub
Hugging Face
LinkedIn
X
HydroLAKES
50,000 lake polygons (of ~1.4 million globally) with attributes, validated geometries, standardized area/perimeter fields and coordinate system.
Source & Citation
Messager, M.L., Lehner, B., Grill, G., Nedeva, I., Schmitt, O. (2016). Estimating the volume and age of water stored in global lakes using a geo-statistical approach. Nature Communications, 7, 13603.… See the full description on the dataset page: https://huggingface.co/datasets/NoraResearchLab/hydrolakes.gedtm30
Quick Links
Website
GitHub
Hugging Face
LinkedIn
X
GEDTM30
2 of 2 global 30 m Cloud-Optimized GeoTIFF terrain tiles, validated (CRS, nodata, pixel dimensions, elevation range via windowed reads) and indexed. Terrain derivative generation (slope/aspect/ruggedness) is left to downstream use of the COGs directly, per the source's own tiled workflow design.
Source & Citation
Ho, Y., et al. (2025). GEDTM30: global ensemble digital… See the full description on the dataset page: https://huggingface.co/datasets/NoraResearchLab/gedtm30.hydroatlas
Quick Links
Website
GitHub
Hugging Face
LinkedIn
X
HydroATLAS
8 sub-basin polygons (level 1 sample) and 50,000 river reaches (sampled), modernized to GeoParquet from HydroATLAS v1.0. This 10-dataset pipeline run uses a sampled subset for memory/time reasons -- see the dedicated HydroATLAS notebook for the full 12-level, 281-attribute treatment.
Source & Citation
Linke, S., Lehner, B., Ouellet Dallaire, C. et al. (2019).… See the full description on the dataset page: https://huggingface.co/datasets/NoraResearchLab/hydroatlas.