A modeling team in the organization using Sentinel-2 (S2) data spatially aggregated at the farm field level in their model wanted to acquire the same features in the same farm fields but in previous years in which field boundaries were missing.
S2 data aggregated spatially by unique crop type derived from the CDL layer within five hundred thousand field boundaries for five historical years along with several observation-level quality metrics (cloud coverage, crop group acreage, crop group shape).
Descartes Labs (DL) Python APIs to retrieve and mask remote sensing data at scale; spatial aggregation and quality metrics calculation with numpy and xarray, AWS CLI to upload very large quantities of data.
Remote sensing and geospatial pipeline for retrieving imagery from S2 and CDL, masking the S2 image by the three most prevalent crops in the field, aggregating the S2 data in each masked region of the field, calculating quality metrics for each crop mask, and storing aggregated S2 data on AWS S3.
