a natural gas fundamentals model that runs in the browser

Built a fundamentals model for AECO, Alberta’s natural gas hub, that runs entirely in the browser. You move the inputs, the Python recomputes, and the price, basis, and trade signals update live. There is no backend. The whole model compiles to WebAssembly and executes on the page.

what it models

You set the drivers: base production in Bcf/d, DUC inventory, LNG utilization, weather, storage, condensate price, and the global price stack of Henry Hub, JKM, TTF, and Station 2 basis. The model returns an AECO price in C$/GJ, a Q1 and annual basis forecast, and a storage range for the year.

The interesting part is what sits underneath the price. AECO does not clear on gas supply and demand alone. It clears on liquids.

the liquids insight

The model makes the condensate link explicit. At C$58/bbl condensate and a 50 bbl/MMcf NGL yield, the liquids uplift is around $2.90/MMBtu. That uplift keeps DUC completions running even when AECO basis is deeply negative.

This is the reason a basis of negative $2.30 in July did not shut in supply. Operators were chasing the condensate barrel and taking the gas as a byproduct. A model that treats AECO as a gas-only market misses this every time.

LNG netback and trade signals

A waterfall walks JKM back to the wellhead: subtract shipping, liquefaction, the Coastal GasLink toll, and regas to get a netback, then compare it against the AECO forward. The model also runs a simple discipline check. It compares its own basis against the forward basis and only flags a trade when the gap clears a threshold, so the tool tells you when there is nothing to do.

scenarios

The output ends in a probability-weighted scenario table. A base case at C$5.00/GJ with the majority of the weight, then upside and downside branches tied to the drivers that actually move the hub: LNG Canada ramp, storage, and weather.

why it works this way

Energy desks live in spreadsheets that nobody else can open or trust. Putting a real model on a web page changes that. Anyone can pull the levers, see the causal chain from condensate to completions to basis, and pressure-test the logic without installing anything. The calibration here is approximate and the tool is a prototype for discussion, but the structure is the point: the fundamentals, the liquids economics, and the LNG arbitrage in one place, transparent and interactive.