Abstract: For almost a century, map makers have looked for a way to construct "cartograms," maps in which the sizes of regions are proportional not to their area on the ground but to some other quantity, such as human population. Cartograms are useful for the visualization of many types of human data, such as disease incidence, census results, and election returns. Unfortunately, the construction of cartograms turns out to be a challenging undertaking -- the obvious approach, of looking for a density-equalizing conformal mapping, turns out to be unsolvable in general. This talk describes a new computer cartogram algorithm using k-space methods that generates elegant and readable maps. A number of applications will be discussed including the much publicized "purple map" of the US presidential election results.

Refreshments at 4:30 in 657 Rhodes Hall.