About

Earth Observation & Geospatial Data Science

I turn open satellite data into validated information: drought status for a vineyard, ground motion checked against GNSS and levelling, and the pipelines that produce them. Based in Poland.

Background

My work starts from a question someone actually needs answered, for example "is this field entering drought?" or "can this InSAR product be trusted for subsidence monitoring?". Then I build the pipeline and test it against independent ground measurements.

I fix acceptance criteria before computing results, I report every number with its error, and I publish negative results too. A layer that has not been validated is labelled as such.

Most of my tools are open: Copernicus data (Sentinel-1/2, ERA5-Land, EGMS), Python, Google Earth Engine and reproducible notebooks.

Technical Stack

Satellite Data

Sentinel-1Sentinel-2ERA5-LandEGMSLandsat 8Airborne LiDARGoogle Earth Engine

Analysis & Validation

Time-series anomaliesInSAR validationIn-situ match-upsBootstrap CIsBrier scorePre-registered criteria

Engineering

PythonpandasrasterioGeoPandasPostGISQGISArcGIS Pro

Machine Learning

scikit-learnTensorFlowPyTorchSuper-resolution (SEN2SR)

Learning Now

STACxarrayDockerGitHub ActionsU-Net

Open to conversations

About Earth observation, validation of satellite products and geospatial data science.

Get in Touch