Leontief functions are a list of what inputs and outputs each technology a country uses as inputs and yields as outputs. So farming a region might require 100 tractors (inputs), of which 95 will be in good condition at the year's end (outputs), and need 50 000 hours of labourer time (input) and yield 10 000 tons of corn (output).
Then a network of trade will usually give you circular graphs: Country A produces corn that is imported by country B to feed its workers. Country B produces iron that is imported by country C as an industrial raw material. Country C produces tractors that are imported by country A for farming.
I think Leontief functions are one of the cooler ideas in economics: they get very close to the empirical data and they allow you to leverage linear algebra effectively. They've been important in some fundamental debates in economics, particularly the Cambridge capital controversy.
I don't know that work done by a famous, and Nobel prize-winning, author really qualifies as "tracking down."
This is a superficial article and I have not read the underlying paper, but in the article it is unclear how wide was the literature search.
>One important question is: what is the value of each sector when they are so tightly integrated? Leontief’s answer was to develop an iterative method of valuing each sector based on the importance of the sectors that supply it.
That's actually pretty interesting. But I think in economics a circular graph doesn't really make sense, does it? It's more two ways with one person trading with another person (with any number of intermediate nodes.)