datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
solar-map-ph
SolarMap.PH data products
Open rooftop-solar detections for the Philippines, built from satellite imagery by SolarMap.PH. Rooftops are detected with CLIP image embeddings plus a gradient-boosted classifier, then segmented and snapped to OpenStreetMap buildings. All geometry is EPSG:4326 (WGS84). CC-BY-4.0.
The trained model itself lives in a separate repo: xmpuspus/solar-map-ph-clf-v4.
Coverage
Per-building and per-tile detections across NCR plus Cebu, Davao… See the full description on the dataset page: https://huggingface.co/datasets/xmpuspus/solar-map-ph.leaves-ph
Leaves.PH: Metro Manila tree-canopy series
Open per-LGU and per-barangay annual tree-canopy estimates for the 17 LGUs of
Metro Manila (2019 to 2026), measured from Sentinel-2 imagery with a
human-calibrated classifier. Canopy here is a dense-vegetation proxy, not a
pixel-exact tree census. Read each year as a cross-sectional snapshot, not a
validated change series.
What is in this repo
per_lgu/per_lgu_canopy_2019_2026.csv… See the full description on the dataset page: https://huggingface.co/datasets/xmpuspus/leaves-ph.