This tutorial discusses Earth Observation (EO) data cubes and how they can be created and processed with the gdalcubes R package. In the first part, the package and its basic concepts (image collections, lazy evaluation, chunking, parallelization) are introduced using real world satellite-based Earth observations from Landsat. The second part will focus on the creation and processing of multi-variable data cubes, containing observations from different satellite-based EO missions. As such, we will look at two study cases on (i) combined vegetation monitoring with Sentinel-2, Landsat. and MODIS data, and (ii) combining vegetation observations with environmental variables (precipitation, soil moisture). Practical challenges, limitations, and ongoing work will be discussed. The tutorial will close with a short practical part, where participants can work on exercises using a smaller dataset.