Skip to main content
Work in Progress

This page is currently under development. Content may be incomplete or contain inaccuracies. If you notice any errors or have suggestions, please contact us.

Quick Start by Role

PhD students & Postdocs — Start with Data Types and Use Cases. For cloud platforms, see Terrabyte and Infrastructure.

PIs & Group Leaders — Review Metadata Standards (ISO 19115, INSPIRE) and Repositories. See also Data Management Plans.

Data Stewards — Key cross-references: File Formats, Repositories, Infrastructure.

What is Geospatial & Earth Observation Data Science?

Geospatial data science deals with data that has a spatial component — information tied to locations on the Earth's surface. Earth observation (EO) extends this by leveraging satellite, airborne, and ground-based sensors to systematically monitor the planet. Together, these fields combine geography, remote sensing, computer science, and statistics to extract knowledge from spatially referenced datasets.

The Need for Geospatial Data Management​

Geospatial and EO data present unique data management challenges:

  • Volume — A single Sentinel-2 satellite scene is ~1 GB; global archives reach petabytes
  • Coordinate reference systems (CRS) — Data must be consistently projected and georeferenced
  • Temporal dimension — Many analyses require time series spanning years or decades
  • Heterogeneous sources — Combining raster imagery, vector boundaries, point measurements, and tabular attributes
  • Resolution trade-offs — Spatial, temporal, and spectral resolution vary across sensors and must be harmonized

Key Application Areas​

Geospatial and EO methods are applied across many disciplines:

  • Environmental monitoring — Land use/land cover change, deforestation, glacier retreat
  • Climate science — Temperature trends, precipitation patterns, carbon flux estimation
  • Urban planning — Infrastructure mapping, population density, mobility analysis
  • Agriculture — Crop monitoring, yield prediction, precision farming
  • Disaster management — Flood mapping, wildfire detection, earthquake damage assessment
  • Ecology and biodiversity — Habitat mapping, species distribution modeling

Common Tools and Software​

GIS Platforms​

  • QGIS — Open-source desktop GIS for visualization and analysis
  • ArcGIS — Commercial GIS platform (ESRI), widely used in industry and academia
  • Google Earth Engine — Cloud-based platform for planetary-scale EO analysis

Programming Libraries​

  • GDAL/OGR — The foundational library for reading/writing raster and vector geospatial formats
  • Rasterio (Python) — Pythonic interface for raster data
  • GeoPandas (Python) — Extends pandas with spatial operations
  • sf (R) — Simple features for R, modern spatial data handling
  • terra (R) — Raster and vector analysis
  • xarray (Python) — Multi-dimensional labeled arrays, ideal for NetCDF/climate data

Remote Sensing​

  • SNAP (ESA) — Sentinel Application Platform for satellite data processing
  • Orfeo ToolBox — Open-source image processing for remote sensing
  • ENVI — Commercial remote sensing analysis software

Infrastructure and Support​

Researchers working with geospatial data at TUM can access:

  • LRZ Linux Cluster — For large-scale raster processing and modeling
  • Terrabyte — LRZ platform specifically designed for Earth observation data science
  • Copernicus Data Space — Free access to Sentinel satellite data
  • Google Earth Engine — Cloud computing for global-scale analysis

Getting Started​

If you're new to geospatial data science:

  1. Explore our resources — Check out the Data Types and Use Cases
  2. Access infrastructure — Apply for Terrabyte access at LRZ
  3. Get support — Contact rdhub@mdsi.tum.de for data management guidance