Do you think researchers who dedicate their lives to measuring poverty haven't thought about a problem you thought of in a few minutes?
> These poverty figures take into account non-monetary forms of income — for poor families today and in the past, this is important, as many of them are subsistence farmers who live largely from their own food production.
This dataset certainly does not do this adequately. This is a very active area of research, and it will take a long time before we converge to reasonable estimates of the past. Even today, countries routinely change their GDP figures after a few years because of how difficult such a thing is to to measure.
You should mentally put large error bars on these graphs. The fact that they don't come up with them already is concerning.
> These poverty figures take into account non-monetary forms of income — for poor families today and in the past, this is important, as many of them are subsistence farmers who live largely from their own food production.