World Seaborne Trade in Real Time: A Proof of Concept for Building AIS-based Nowcasts from Scratch
IMF Working Papers, May 14, 2020
Source details
- Canonical URL
- World Seaborne Trade in Real Time: A Proof of Concept for Building AIS-based Nowcasts from Scratch
Other formats
Bibliographic details
- Authors: Diego A. Cerdeiro, Andras Komaromi, Yang Liu, Mamoon Saeed
- Published: May 14, 2020
- Series: IMF Working Papers
- DOI: https://doi.org/10.5089/9781513544106.001
Summary and main findings
- Authors: Diego A. Cerdeiro, Andras Komaromi, Yang Liu, Mamoon Saeed
- Publication date: May 14, 2020
- Core finding: AIS (Automatic Identification System) radio-signal data can be transformed into economically meaningful, policy-relevant indicators of international trade using an end-to-end algorithmic pipeline.
- Constructed indicators:
- World-level seaborne trade indicators.
- Bilateral trade indicators.
- Within-country trade indicators.
- Fit with official statistics: The methodology "achieves a good fit with official trade statistics for many countries and for the world in aggregate."
- Use cases demonstrated:
- Sectoral analyses of crude oil trade.
- Event studies, including Hurricane Maria and measures to contain the spread of the novel coronavirus.
Methodology and technical approach
- Data source: Raw AIS messages emitted by the global vessel fleet for navigational safety purposes.
- Key algorithmic components:
- Machine-learning techniques to identify port boundaries.
- Algorithms to construct port-to-port voyages from AIS tracks.
- Estimation routines to convert voyage information into trade volumes.
- Scope: A globally applicable, end-to-end proof of concept built from scratch using raw AIS signals.
Validation, applications, and examples
- Validation: Comparison with official trade statistics at country and global levels; reported as a good fit for many countries and in aggregate.
- Demonstrated applications:
- Crude oil trade sectoral analysis.
- Event analysis: Hurricane Maria.
- Event analysis: effects of measures to contain the spread of the novel coronavirus.
Limitations and recommendations for improvement
- Identified limitations and paths for improvement:
- Ongoing refinements of the algorithms are needed.
- Additional data on vessel characteristics would improve estimates.
- Incorporating country-specific knowledge should help performance for several country cases.
Publication and metadata
- Series: Working Paper No. 2020/057
- Issue: 057
- Volume: 2020
- Pages: 44
- DOI: https://doi.org/10.5089/9781513544106.001
- Stock No: WPIEA2020057
- ISBN: 9781513544106
- ISSN: 1018-5941
- Citation preview (format: Chicago): Diego A. Cerdeiro, Andras Komaromi, Yang Liu, and Mamoon Saeed. "World Seaborne Trade in Real Time: A Proof of Concept for Building AIS-based Nowcasts from Scratch", IMF Working Papers 2020, 057 (2020), accessed 9/18/2026, https://doi.org/10.5089/9781513544106.001
IMF Working Paper (May 14, 2020) — Summary and metadata drawn from the publication landing page.
Content in this bundle
- Working Paper