Bayesian Probabilistic Population Projections: Do It Yourself

Hana Sevcikova, University of Washington
Patrick Gerland, United Nations Population Division

A Bayesian approach for probabilistic population projections has recently been used by the United Nation Population Division in the preparation of the 2012 revision of the World Population Prospects. The methods have been implemented in publicly available open-source software as a collection of R packages. In this paper, we demonstrate how to easily reproduce such population projections, including probabilistic projections of total fertility rate and life expectancy. The packages allow any analysts to generate variations of the UN projections, to use their own data, to impute missing data or to apply the methods to sub-national datasets. Using a flexible expression language, probabilistic results can be summarized and visualized in graphs, maps, or population pyramids. The software can be conveniently controlled from a graphical user interface.