Shock Cascade

The Problem

Macro risk is hard to reason about because disruptions don't stay in one lane. A supply chain shock in semiconductors ripples into auto manufacturing, consumer electronics, and eventually retail earnings. Existing tools show exposure to individual sectors but not the cascade path between them.

The Approach

Built a causal graph simulator where users model a disruption (tariff, supply shortage, rate hike) and trace its cascade effects through interconnected sectors. The graph maps first-order, second-order, and third-order impacts, then overlays exposure onto a sample portfolio.

Key decisions:

• No AI/ML: the causal relationships are hand-mapped and transparent, not black-box predictions
• Used D3.js force-directed graphs to make cascade paths visually traceable
• Portfolio overlay shows dollar-weighted exposure at each cascade level
• Deliberately kept the model simple enough that users can challenge the assumptions

The Outcome

Prototype maps 12 major sectors with 40+ causal links. Users can simulate scenarios like "what happens to my portfolio if Taiwan chip production drops 30%" and see the cascade unfold step by step. The visualization makes second-order effects obvious in a way that spreadsheet models don't.

Reflection

The biggest design decision was choosing transparency over sophistication. An ML model could predict more nuanced cascades, but you can't argue with it or learn from it. The hand-mapped graph forces users (and me) to think about WHY a disruption propagates, not just that it does. If I expanded this, I'd add historical scenario backtesting to validate the causal weights.