Instructor Notes

Instructor notes


Episode 3

  • brpgewaspercelen_definitief_2020_small.gpkg was created because the original file was too large to download and load. Original file, which was ~500Mb could take several minutes to load, and could crash the Jupyter terminal.
  • The cropped version of brpgewaspercelen_definitief_2020_small.gpkg: data/fields_cropped.shp is required for later episodes.
  • The “Modify the geometry of a GeoDataFrame” section is optional and can be skipped without consequences.

Episode 4

  • It is not recommended to plot the visual band directly, due to its size (3 x 10980 x 10980). Please plot the overview as in the teaching material.
  • The clipped raster data: raster_clip.tif is required for later episodes.

Episode 5

  • The calculated NDVI: NDVI.tif is required for later episodes.
  • The calculated classification identifier: NDVI_classified.tif is required for later episodes.

Workshop setup


  • Consider using mamba for speeding up the Python environment setup.
  • Make sure the audience has downloaded the three vector datasets to the data repository.

Introduction to Raster Data


Introduction to Vector Data


Coordinate Reference Systems


The Geospatial Landscape


Access satellite imagery using Python


Considerations for the position of this episode in the workshop

When this workshop is taught to learners with limited prior knowledge of Python, it might be better to place this episode after episode 11 and before episode 12. This episode contains an introduction to working with APIs and dictionaries, which can be perceived as challenging by some learners. Another consideration for placing this episode later in the workshop is when it is taught to learners with prior GIS knowledge who want to perform GIS-like operations with data they have already collected or for learners interested in working with raster data but less interested in satellite images.



Extra attention for the following exercise

  • The exercise Exercise: Search satellite scenes using metadata filters needs extra attention. Its output search.json is required for the later episodes. Therefore we recommend:
    • Do not skip this exercise;
    • Think twice when you would like to change the query arguments in this exercise;
    • Make sure all the audience have the output search.json before continuing.


Read and visualize raster dataResampling the raster image


Vector data in Python


Crop raster data with rioxarray and geopandas


Raster Calculations in Python


Calculating Zonal Statistics on Rasters


Parallel raster computations using Dask


Data cubes with ODC-STAC