Modelling and Smoothing Mortality Surfaces and developing the related R package: MortalitySmooth

Aggregate data on mortality over age and time, so-called mortality surfaces, are one of the most important sources of information for the study of mortality trends. Much can be gained from visual displays of these surfaces, but deeper insights can only be obtained by modeling these surfaces. Building on my PhD research, I further developed P-spline models for the analysis of mortality surfaces. These models provide two-dimensional smoothing techniques for Poisson counts, which are flexible enough to cover the trend patterns found in mortality development. To make this technology accessible to statistically less experienced researches, I developed the software package (MortalitySmooth) for use with the free and open-source statistical software R. It can be downloaded free of charge and you can find a special page about here.

In 2012, I published a related article of the Journal of Statistical Software in which I gave on overview of the package, its methodology behind and several examples about its usage: MortalitySmooth: An R Package for Smoothing Poisson Counts with P-Splines (link).