The above map shows the sites of recent and upcoming meetings sponsored by SPES, overlaid on a map that estimates the population density of SPES and Q&P members.
The location of each member is estimated from the zip code of the member's ASA mailing address. A bivariate kernel density estimator with a bandwidth multiplier of 0.5 is used to estimate the population density of members per square unit. The units are latitude and longitude; since the number of square miles in a one-degree latitude by one-degree longitude patch of earth varies according to the latitude of the patch, this population density is not exactly interpretable in terms of square distance. However, a rough approximation is that the density estimate smooths individuals over approximately a 30 square mile area.
Each county in the contiguous USA was colored according to the density estimate scored at the county's geometric centroid. Thus the counties create a piecewise constant approximation to the density estimate.
The graphic was created by Rick Wicklin of SAS Institute, Inc., using SAS® Stat Studio 3.1 (www.sas.com/statistics).