It’s Friday the 13th. So why not a morbid post? I stumbled onto to some “life tables” for 2007. Actually what the link gives you is the probability of death and life expectancy at a given age for males and females. This is obviously important for purposes of private and public insurance. The tables are naturally broken down by sex because males and females differ in their likelihood of death. Women are well known to have higher life expectancies in all but the most primitive of nations today.
I wanted to do a little visualization. I limited the years across the interval 0 to 100. Since I find life expectancy a boring statistic (well known) I want to focus on probability of death in a given year. I generated four tables.
– First, a table which shows probability of death in a given year log transformed. I had to transform because the difference between a 1 year old and a 100 year old spans four orders of magnitude.
– Second, a table which illustrates the absolute change in the probability of death year to year. This is basically measuring the absolute change between year X and year X – 1. Concretely an …