Current Topics in Demographic Research
Fall 2010

Project 1 The Human Mortality Database: A joint project between UC Berkeley and the Max Planck Institute for Demographic Research, the HMD holds a wealth of great information on human longevity. The data are carefully constructed to allow cross country comparisons. For some countries data are available as far back as the 18th Century. In this project we explore the structure of the database and some of the uses of these data. |  download:
Project 2 Endogamy and cousin diversity
One consequence of intergroup marriage is diversity among cousins. The prevalence of marrying outside of ones group is the key determinant the diversity of cousin sets. In perfectly endogamous societies, all cousins must be of the same group. In this project we will use Socsim to investigate the changing patterns of group membership in cousin sets as patterns of inter group marriage change over time. 		|  download zipped tar file
Project 3 Lee Carter Forecasting The "Lee-Carter" forecasting technique is a staple for Demographers. Originally used for forecasting mortality, modifications and extensions have made the procedure useful for host of other applications. Generally, LC is useful for forecasting and interpolating sets of values that are correlated across some index like age and time. |  download mortmodel1.R |  download LCintro.R
Project 4 Diffusion cartograms and tests of complete spatial randomness. The diffusion cartogram algorithm developed by Gastner and Newman (See 2004 PNAS article) Provides a clever way of modifying the boundaries of regions to represent some quantity. While occasionally useful for creating thematic maps, it is even more valuable as a way of controlling for population density (or some other quantity that varies geographically) in order to test for spatial randomness. A classic use would be to adjust locations of observed cancer cases by the the density of the at risk population. If the resulting locations are not spatially random then a case could be made for designating a cancer cluster. We will first apply the cartogram procedure to test for the randomness in the location of fastfood restaurants. Then we will adapt the code in order to test other location data. |  Instructions and code