Extendible metric MDS in Python (akin to landmark MDS). Classic metric MDS transforms a symmetric matrix of metric distances between N objects into a set of projections in an orthonormal space defined by an eigendecompisition of the distance matrix (after some transformations). Thus, the method is akin to PCA, though unlike PCA it doesn't provide a straightforward way to project new objects onto the space (i.e. a projection matrix). This package implements an active metric MDS method described in "Principal Component and Correspondence Analysis in R" by Dr. Herve Abdi (2017) and is thus similar to package bios2mds written in the R language. The method allows new objects to be projected onto an existing space defined by a set of active objects.