Bayesian Demographer: Software for Probabilistic Population Projections

• Activity Number: 409
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Computing
• Author(s): Hana Sevcikova*+ and Adrian Raftery and Leontine Alkema
• Companies: University of Washington and University of Washington and National University of Singapore
• Keywords: Bayesian hierarchical model ; Markov chain Monte Carlo ; fertility projection ; life expectancy ; population projection ; United Nations

Abstract

Every two years, the United Nations Population Division (UNPD) publishes projections of the populations of all countries of the world, called the World Population Prospects. Instead of assessing uncertainty about the projections, the UNPD produces High, Medium and Low variants of these. However, these are scenarios and lack a probabilistic interpretation. We have now developed new methods for probabilistic projection of the total fertility rate (TFR) and life expectancy for all countries, and the UNPD is planning to use them as inputs to their 2010 projections (to be published in 2011), as a first step towards fully probabilistic projections in the future. To support the UNPD in this effort, we have developed several R packages, that implement probabilistic projections of TFR (bayesTFR) and life expectancy (bayesLife), as well as an easy to use graphical user interface (bayesDem) that combines these components into one convenient bundle. In this talk, a brief background of the projection methods will be given. Then the structure of the software components will be discussed, and the software will be demonstrated.