In the first part of this paper we describe the Cartographic Structure-Matching (CSM) algorithm that identifies multi-object structures from topological data. We describe the significant differences between CSM, and traditional structure matching developed for the purposes of cognitive modelling. CSM was initially developed as a tool for categorising topological data based on the context of some unclassified object. (Traditional techniques examine only an objects content, like size and shape). CSM looks at the categorisation of objects adjacent to the unclassified object, and uses this as a basis for classification. CSM is particularly adept at classifying objects whose shape does not uniquely identify it. The remainder of this paper describes a number of more recent applications of CSM.