Commodities models

Some of the others have touched on government templates and proprietary data sources, so I'll try to elaborate. The models themselves are pretty straightforward in that they are supply sources minus demand sources with the balance flowing into inventories. Given data is not uniform will generally organize supply and demand into "buckets" that are easier to track and can be forecast. Similarly the inventories data will be checked against available sources with various "adjustment factors." Knowing where to find data sources with some predictive power, where to track realized outturns, and how one should treat "adjustment factors" is the hard part, with the even harder part being translating that to price forecasts. If excel/accounting are the tools for fundamental analysts then r/python/excel/econometrics/statistics/data analytics are the tools for commodity analysts to answer those questions I referenced (of course supplemented with some commodity specific knowledge).

I’ll add some examples for context- imagine trying to project the recovery in miles driven from the height of the pandemic as a proxy for oil demand for autos and thus your demand bucket. Similarly imagine trying to figure out maximum sustainable production capacity for Saudi Arabia to get a sense for how much oil could be supplied, and thus impact your balances/price forecasts, in a price war. Or on the inventory side trying to determine how much China is importing in excess of needs (perhaps for strategic reasons) and whether that will continue.