In this project, we perform data-driven analysis and modeling of international trade, using the historical trade data between nations. Our approach combines dynamical systems identification with graph theory to capture both the temporal evolution of trade flows and the network of interdependencies they form. Emphasis is placed on interpretability: rather than building opaque predictive models, we aim to construct representations that can be understood and related to the economical and political data. First of all, we adapt statistical dynamics identification methods to explicitly incorporate the network structure, and we will calibrate them using synthetic data. Secondly, we apply these methods to the data collected by the World Trade Organization in order to extract meaningful patterns.

Through this SuperUROP, I want to apply my skills in an interdisciplinary project that takes a mathematical approach to an economics based problem. This project lies in the intersection of my two majors, 18 (Math) and 6-14 (CS, Econ, and Data Science). Having taken classes on networks (6.326), dynamical systems (18.353), and high dimensional statistics (18.656), I’m excited to apply the knowledge I gained from these classes to real world data.