CGC: a Scalable Python Package for Co- and Tri-Clustering of Geodata Cubes

Abstract

Multidimensional data cubes are increasingly ubiquitous, in particular in the geosciences. Clustering techniques encompassing their full dimensionality are necessary to identify patterns "hidden" within these cubes. Clustering Geodata Cubes (CGC) is a Python package designed for partitional clustering, which identifies groups of similar data across two (e.g., spatial and temporal) or three (e.g., spatial, temporal, and thematic) dimensions. CGC provides efficient and scalable co-and tri-clustering functionality appropriate to analyze both small and large datasets as well as a cluster refinement functionality that supports users in their quest to make sense of complex datasets.