Tuesday, April 20, Volen 101, 3:00 p.m.
Exploratory data analysis is critical in a broad range of research areas, where large collections of multidimensional data need to be meaningfully arranged and presented. Sorting Points Into Neighborhoods (SPIN) is a novel method for the organization and visualization of data, implemented in a simple interactive tool. SPIN utilizes traits of distance matrices to find a natural ordering that highlights the underlying shape of the original, multidimensional data. Toy models will show how structures and the relationships between them can be inferred from the reordered distance matrix. Examples from diverse disciplines, from cancer-research to computer vision, will demonstrate that these structures have meaningful interpretations.
Host: Jordan Pollack