Skip to content

Tutorials

Step-by-step notebooks that build models with EconIAC from scratch. Each tutorial starts from a concrete economic problem and introduces library concepts as they are needed.

To run locally:

pip install "econiac[tutorials]"
jupyter notebook docs/tutorials/keen_predator_prey.ipynb

To run in Colab: click the notebook link, then uncomment the !pip install econiac jax[cpu] cell at the top.

Available tutorials

Notebook Concepts introduced Audience
MONIAC — the hydraulic economy, differentiable Phillips (1950) ODE, PCL conservation, Gibbs slot-cam, jax.grad fiscal multiplier, accelerator bifurcation, susceptibility χ early-warning Everyone — start here
Budget simplex and the demand law Budget simplex geometry, gibbs_weights β-sweep, β* calibration from choice variance, Smith information transfer index = log Z_β Economists, theorists
Keen predator-prey ODE simulation, BalanceSheet, PCL choose, TIR routing, thermal Shapley, conservation_loss Economists, macro modellers
GEMMES: Keen + climate CurvedBalanceSheet, stranded assets, PCL fold, 4-player Shapley, carbon tax routing Climate economists, central bankers
GL Model PC: portfolio choice gibbs_weights, TIR portfolio allocation, calibrate_beta, β* from Flow of Funds data, phase boundary at β=0 Monetary economists, SFC modellers
LowGrow SSE: green transition CurvedBalanceSheet, climate damage curvature, TIR investment routing, calibrate_green_beta, carbon tax phase diagram, carbon lock-in paradox Climate economists, energy modellers
Cross-currency swap: gauge theory FXMarket, YieldCurve, CIP holonomy, jax.grad Greeks, cross-gamma Hessian, 3-currency curvature Quantitative analysts, risk managers, derivatives practitioners
Supply chain RST: differentiable ADTs SupplyCapacity (AND/Product type), FinancialRisk (OR/Sum type), reverse_stress_test, laplacian_spectrum, PCL combinator duality, Curry-Howard correspondence Supply chain analysts, risk managers
Fraud detection via topology BalanceSheet, Connection, wilson_loop, curvature_matrix, PCL type-safety, jax.grad equity sensitivities, Wilson loop geometric audit Accountants, auditors, financial regulators

Why notebooks, not scripts?

The .py scripts in examples/ are the same computations run straight through. The notebooks add:

  • Prose explaining the economic interpretation of each step
  • Plots showing trajectories, phase portraits, routing weights, and attribution bars
  • Colab compatibility — one cell to install, the rest runs in the browser

The examples/ scripts are the canonical reference; the notebooks are the tutorial layer.