isp-uv-es/usavars-taco

USAVars

This is a repackaging, not a new dataset. It is USAVars by UC Berkeley (Rolf et al., MOSAIKS), USDA NAIP, converted to TACO. Pixel values and labels are kept as released except where the description below says otherwise. All credit belongs to the original authors: if you use it, please cite them and follow their licence.

original dataset · paper · licence: CC-BY-4.0

Repackaged into TACO by the Image and Signal Processing Group (ISP), Universitat de València, within the ELLIOT project.

Citation

Please cite the original work:

@article{rolf2021mosaiks,
  title   = {A generalizable and accessible approach to machine learning with global satellite imagery},
  author  = {Rolf, Esther and Proctor, Jonathan and Carleton, Tamma and Bolliger, Ian and Shankar, Vaishaal and Ishihara, Miyabi and Recht, Benjamin and Hsiang, Solomon},
  journal = {Nature Communications},
  volume  = {12},
  number  = {1},
  pages   = {4392},
  year    = {2021},
  doi     = {10.1038/s41467-021-24638-z}
}

About the data

97876 NAIP aerial tiles (4 bands R/G/B/NIR, 4 m, 256x256 px = ~1 km2) sampled uniformly at random across the contiguous United States.

97,876 samples · splits: test 9,788 · train 68,513 · validation 19,575 · tasks: regression

Packaged as TACO v3.

Full description

Target. Scalar regression on three continuous variables per tile: tree cover, elevation and population density.

Splits. The official train/val/test split is preserved.

Missing values. The release's -999 sentinel is stored as NaN, with a per-target validity flag.

Scope. The release's other four label tables (housing, income, roads, nightlights) are keyed to the POP sample, whose imagery is not part of this release, and are not carried.

Getting started

git clone https://github.com/OscarPellicer/taco
pip install -e "taco/python[ml]"     # builds the native reader: needs CMake and Ninja

Read it straight from the Hub:

from huggingface_hub import snapshot_download
from taco.ml import Dataset, plot_sample

path = snapshot_download("isp-uv-es/usavars-taco", repo_type="dataset")
ds = Dataset(path)                  # the repository is the container: parts + .tacocat
plot_sample(ds[0])

or from a local copy:

ds = Dataset("usavars")
sample = ds[0]                      # {slot name: SlotValue}, arrays decoded
sample["image"].array.shape

Metadata without decoding anything:

import taco
taco.read("usavars/.tacocat")              # one Arrow table, levels joined

Samples

What a sample contains

role	slot	holds	modality	detail
input	image	raster	optical	4 band(s), requantised
target	treecover	scalar		unit %
target	elevation	scalar		unit m
target	population	scalar		unit count/km2

Licence

CC-BY-4.0

Terms of use:

• Attribution required (CC-BY-4.0).

Required credits:

• USAVars: Rolf et al. (MOSAIKS); TorchGeo
• Imagery: USDA NAIP (public domain)
• Tree cover: Hansen/UMD/Google/USGS/NASA Global Forest Change
• Population: CIESIN Gridded Population of the World v4
• Elevation: Mapzen terrain tiles

Providers: UC Berkeley (Rolf et al., MOSAIKS), USDA NAIP

Acknowledgements

TACO was designed by César Aybar and is specified at https://asterisk.coop/taco/spec/.

Built by Oscar Pellicer within the ELLIOT project at the Image and Signal Processing Group (ISP), Universitat de València.