CASE STUDY · 2026
Environmental Data API
A production Earth-observation data service for wildfire-risk research.
The service is represented conceptually because its endpoints, infrastructure and operational details remain private.
- CONTEXT
- Wildfire-risk research in Portugal
- ROLE
- Main developer
- WHEN
- 2026
- STACK
- FastAPI · Rasterio · STAC · Nginx
The service
This production FastAPI service supplies environmental and Earth-observation data to an external team developing a wildfire-risk prediction model for Portugal. It integrates providers with different APIs, formats and processing requirements behind a reusable interface.
What I owned
- Designed the service and implemented the Landsat 9 workflow, from authenticated search and pagination through processing, caching and packaging.
- Implemented CRS-aware cropping, QA masking, surface-temperature conversion and STAC metadata.
- Built the ERA5-Land integration and a NASA Earthdata VIIRS HDF5-to-Parquet pipeline.
- Integrated pre-existing Copernicus retrieval work from another developer without attributing its algorithms to myself.
- Handled most production deployment work, including later Nginx and API-key middleware updates.
Engineering highlights
Multi-provider integration
Unified Copernicus, ERA5-Land, USGS Landsat and NASA Earthdata despite different access patterns and formats.
Geospatial processing
Implemented CRS-aware raster operations, quality masking and scientific-data transformations.
Reusable metadata
Introduced STAC metadata so processed outputs remained consistent and discoverable.
Real-world use
The production service provides processed data to an external research team working on wildfire-risk prediction for Portugal.