I am always surprised by how many interesting stories a dataset can tell. The latest StatsSA release of the Causes of Death dataset (2012) has a lot to say about death trends in South Africa. We always hear about HIV, TB, maternal and child deaths as leading causes of morbidity in South Africa. I am more interested in the micro-patterns. Here are some anecdotes:
- We often hear that boys are more accident-prone than girls. In the early years, twice as many boys drown as girls. In their 20s, four times as many men drown as women.
- Of the 3,980 gun deaths in 2012, 15 - 40 year-olds accounted for 70% of them.
- Tragically, 8% of all deaths in 2012 involved babies under a year-old.
- We know that smoking is bad for you - what does that really mean? Have a look at the image below:
If you are going to die of cancer, your likelihood peaks at around 62 years. If you are a non-smoker, that likelihood shifts to 65 years. The data includes just under 14,000 cancer deaths. From those, we can see that smokers simply don’t live as long as non-smokers.
In short, these revelations are nothing new. It is still satisfying to see that the data confirms it.
My next project - is it true that married men do actually live longer than their bachelor counterparts?