How I Built a Graph-Based Digital Twin to Simulate Cascading Supply Chain Failures

Building a simulation engine using Python, NetworkX, and Streamlit to model cascading failures in supply chain networks. Most supply chain failures don’t happen all at once — they unfold in cascades. I recently built a graph-based simulation engine to understand how disruptions propagate through complex logistics networks. The system models supply chains as directed weighted graphs and simulates how failures spread step-by-step across infrastructure. In this post, I’ll walk through how I built it, how the cascade logic works, and what I learned from designing the system. Tech Stack Python NetworkX Streamlit Pandas Pytest (for deterministic validation) Modeling the Network The supply chain is modeled as a directed weighted graph: python import networkx as nx G = nx.DiGraph() G.add_node("A", type="factory") G.add_node("B", type="hub") G.add_node("D", type="market") G.add_edge("A", "B", weight=2) G.add_edge("B", "D", weight=2) Routing is computed using Dijkstra’s algorithm: - nx.dijkstra_