We present a new constraint-based automatic cartogram construction method that successfully achieves desired region areas while maintaining map topology and preserving essential shape cues to enable region recognition. When the results are compared with a number of existing methods, they appear to be superior in both accuracy and preservation of shape recognition cues.
A continuous area cartogram is a map transformation in which the map regions are resized relative to the geographic distribution of a data set. By spatially reflecting the data within the map base, the cartogram emphasizes each region's data instead of territorial land area, thereby providing a powerful tool for visualizing data distribution.
There are two distinct and conflicting goals in the construction of cartograms:
adjusting region sizes and retaining region shapes. Our Constraint-Based
Method utilizes three foundational mechanisms to achieve these goals: alternating
relaxation, constrained dynamics, and hierarchical resolution. We
converge upon each goal in an alternating relaxation fashion, by achieving
desired areas without regard to shape, and then utilizing constrained dynamics
to attempt to hold the areas fixed while shape is restored. Through
hierarchical resolution, we perform gross adjustments initially upon a
coarsely resampled map and refinements later at progressively higher levels
Kocmoud, C.J. and D.H. House. "A Constraint-Based Approach to Constructing Continuous Cartograms". To appear in the Eighth International Symposium on Spatial Data Handling Proceedings, Vancouver, BC, July 11-15, 1998.