Data Is Beautiful

I am suspicious of this

So Russia’s death rate was pretty much unchanged from 1930 to 1935 to 1945, and then things got way better in 1950?

Maybe I could see, they are counting only Russia (not the USSR), so the holomodor is largely absent from 1930, and then Russia advances in living standards meant that there was a huge underlying boost that masked the unprecedented deaths during WW2, and then after WW2 the apparent life expectancy shot up because a lot of the vulnerable or old people were already dead. But I don’t buy it. Idk what's going on with their data, but China looks fine and Russia looks simply wrong; it is missing some big dips that it should have.

Edit: Hm, I guess there is a 6-year divot in 1932… I guess I just expected the Holomodor to show up bigger and less spread out over surrounding years. But yeah maybe it is showing up.

Edit 2: Okay, I looked more and I am confident that this isn’t exactly right. It says “the remaining average lifespan for a hypothetical group of people, if they experienced the same age-specific death rates throughout the rest of their lives as the age-specific death rates seen in that particular year.” There’s no possible way that extrapolating out the death rates people were experiencing in the middle of a famine or war would lead to these gentle dips and small divots.

I suspect that by combining data from different sources, they wound up using cohort LEB for the distant past and period LEB for the more recent past. That would explain why e.g. the dip in Russian life expectancy because of the Ukraine war shows as the same size as the dip for WW2. If they were doing the calculations the same for both, the WW2 dip would take away half the chart or more. So maybe it’s not really wrong per se but just mismatching their metrics in a way that makes it hard to draw anything of precision from the chart beyond “things getting better”.