Skill Guide
The systematic process of identifying, acquiring, cleaning, transforming, and integrating spatially-referenced data from diverse sources into an analysis-ready format.
Scenario
Acquire and preprocess satellite imagery and vector land-use data for a selected city to calculate green space per capita.
Scenario
Build a pipeline that periodically acquires precipitation forecasts, elevation models, and river network data to generate potential inundation maps.
Scenario
Design and implement a near-real-time system to detect forest loss using multi-source satellite data, integrating alerts with enterprise GIS.
The fundamental Python/CLI toolkit for reading, writing, and manipulating virtually all geospatial data formats. GDAL is the industry backbone; the others provide more Pythonic interfaces for raster and vector operations.
Provide access to petabytes of pre-processed and raw satellite imagery via scalable APIs and cloud computing environments. Essential for large-area, time-series analysis without managing local storage.
For storage, complex querying, and large-scale distributed processing of vector and raster data. PostGIS is the standard for spatial SQL; Sedona and Dask enable distributed computation on clusters.
Answer Strategy
Demonstrate knowledge of CRS concepts, reprojection methods, and practical tool usage. Sample Answer: 'First, I would use `geopandas` to read the Shapefile and `rasterio` to open the GeoTIFF, inspecting their `.crs` attributes to confirm the mismatch. I would then reproject the vector data to match the raster's UTM CRS using `geopandas.to_crs(epsg=32633)`. For the overlay, I would vectorize the raster's footprint using `rasterio.features.shapes` or use `rasterstats` to extract values directly, ensuring all operations are performed in the same projected coordinate system.'
Answer Strategy
Tests problem-solving, understanding of remote sensing principles, and pipeline robustness. Focus on systematic diagnosis and modular design. Sample Answer: 'I would isolate the failure point by checking logs for specific errors (e.g., band mismatch, projection issues). I would then validate the new data's metadata against our schema. To adapt, I'd implement a configuration-driven preprocessing step where parameters like band order, radiometric calibration coefficients, and mosaicking logic are sourced from a config file per sensor, not hard-coded. This allows the pipeline to ingest new data types by updating configuration, not rewriting code.'
1 career found
Try a different search term.