TariffGraph AI

TariffGraph AI is an AI-powered trade intelligence and visualization platform that helps users understand how tariff changes could affect interconnected trade relationships, products, industries, and downstream businesses. Instead of returning only a text answer, TariffGraph AI converts a natural-language trade question into an interactive Trade Impact Graph. The system identifies relevant entities such as countries, products, tariffs, industries, and time periods, then connects them through trade and impact relationships. The platform separates source-backed facts from AI-generated inference and potential impact. This helps users understand that a tariff change may create cost pressure or downstream effects without presenting uncertain outcomes as guaranteed predictions. Users can explore the graph by clicking and dragging nodes, inspect relationship details, review the reasoning path, examine evidence and trade-data information, view a timeline, and test what-if tariff scenarios such as changing a tariff from 10% to 25%. The prototype uses React, Vite, Tailwind CSS, React Flow, Node.js, Ollama, and IBM Granite locally for AI analysis. IBM Bob 2.0 was used as a core development partner during the hackathon for understanding the repository, implementation, testing, debugging, and maintenance workflows. The project is designed as a prototype with clearly labeled demo trade data where official live datasets are not connected.