Data Cubes
Edzer J Pebesma, Roger Bivand · 2023
Spatial data cubes are array data with array dimensions associated with space dimensions and usually also with time dimensions. Space dimensions can be represented by a regular discretization in x- and y-direction, leading to raster data cubes, or be represented by a set of vector geometries (e.g. points or polygons) leading to vector data cubes. Basic operations on raster and vector data cubes are explained, including filtering (selecting) values for dimensions, applying functions over dimensions (e.g. low-pass filtering), reducing dimensions e.g. by computing the mean or maximum over time, and aggregating dimensions to a lower spatial or temporal resolution, or aggregating raster data over vector geometries. The ability to switch dimensions and attributes leads to very flexible data structures, and powerful options for analysis.