MITGlobalCommodityFlow is a global forecasting model developed by MIT for predicting the flow of imported commodities across multiple major ports in the United States. This project leverages Recurrent Neural Networks (RNNs) to analyze and forecast commodity flows, aiming to enhance the resilience and efficiency of the nationwide supply chain network.
To improve the resilience of the nationwide supply chain network, it is essential to understand the processes at the most upstream part of goods flow in the USA – the ocean ports. By estimating the volume, category, value, and routing of goods, stakeholders can better plan transportation, allocate resources, and manage the ordering and sourcing of goods.
The US port system is a complex network where different commodities are imported through various ports to minimize ground transportation. The project aims to determine if a global forecasting model can effectively predict the weight of imported commodities across ten major US ports.
- Ports: Nine major US ports, including Savannah, New York/New Jersey, Boston, Norfolk, Houston, Los Angeles/Long Beach, Oakland, and Seattle/Tacoma.
- Commodities: Sixteen food commodities categorized using the 2-digit HS code.
- Import Data: Monthly import data from 2003 to 2024, including customs data, containerized vessel value (in USD), and containerized vessel weight (in kg).
We welcome contributions from the community. To contribute:
- Fork the repository.
- Create a new branch:
git checkout -b feature-branch
- Make your changes and commit them:
git commit -m "Description of your changes" - Push to the branch:
git push origin feature-branch
- Create a pull request detailing your changes.
This project is licensed under the MIT License. See the LICENSE file for details.