Figure 7
From: Unveiling patterns of international communities in a global city using mobile phone data
Classification method based on persistent homology. (A) Point clouds defined in metrical spaces can be studied effectively through the features of associated simplicial approximations. In particular, from the actual data sample (top row) it is possible to extract their coarse-grained representations that preserve their homological properties (bottom row). (B) To each recovered shape, we can then associate a pair \((r_{0}, r_{1})\) describing the shape’s spatial coherency and circularity and use this information to classify the original datasets.