We apply digital twin technology to reconstruct complex economic systems, leveraging advanced data science methods to replicate the interconnected network of global trade. We also integrate the transportation network into the digital twin’s architecture to gain holistic visibility into global supply chains. This inductive approach enhances the study of economic competitiveness as machine learning models provide new insights into the evolving nature of markets. It also allows for the identification of new variables that better explain the organization of markets, making economic analysis more effective.
The SLGL dataHub is an in-development platform to access the digital twin of the bi-national St. Lawrence-Great Lakes region. It harmonizes diverse streams of trade and transport data to construct the region’s economic structure and project trade flows within it. By mimicking the characteristics and behaviours of the real-world economy onto the virtual realm, the platform enables simulations and predictions that reveal the micro level effects of macroeconomic phenomena. With this comprehensive and real-time view of the economy, users will gain actionable insights into its performance, efficiency and challenges, empowering them to make informed decisions and strategies.
*
Harmonizing diverse datasets
Geospatial data is foundational both to the SLGL dataHub’s architecture and our subsequent research. We employ an iterative and incremental approach in the development of the SLGL dataHub, gradually integrating datasets to introduce new features and variables. This process harmonizes diverse datasets from the domains of trade and transport, providing high-quality training data for machine learning algorithms used in simulations and predictions with the digital twin. It also links data to geographic features at varying levels of resolution, maximizing the utility of existing datasets.
Researchers also benefit from direct access to the underlying database, which addresses key challenges such as differences in methodologies, observation intervals, classification standards, missing data and formatting.
SLGL dataHub input-output table
Our dual-resolution input-output table serve as the SLGL dataHub’s core dataset. By integrating data from several large datasets, we uniquely reconstruct the structure of global production and trade to simultaneously operate at two geographic levels: the country level and the sub-national level, specifically between Canadian provinces and US states within the SLGL region. This structure makes it possible to trace inter-industry flows between sub-national entities, such as Québec and New York, and also to other countries like Germany.
| Geographic Resolution | Geographic Coverage | Industry Coverage | Time Span | Frequency | Currency |
|---|---|---|---|---|---|
| Country, Sub-national | 43 countries, 2 provinces, 8 states, ROW | 45 Industries | 2012-2020 | Annual | USD |
Data from the following datasets are harmonized and used to reconstruct the SLGL dataHub input-output table.
| Dataset | Source |
|---|---|
| Canadian International Merchandise Trade Web Application | Statistics Canada |
| Freight Analysis Framework, FAF5 | U.S. Department of Transportation, Bureau of Transportation Statistics, Federal Highway Administration |
| Inter-Country Input-Output (ICIO) | OECD |
| StateIO/stateior | U.S. Environmental Protection Agency |
Geospatial datasets
| Dataset | Description | Geographic Coverage | Unit | Source |
|---|---|---|---|---|
| Global Shipping Lanes | Shipping routes geo-referenced from CIA’s Map of the World Oceans | Global | Routes | Benden, P. (2022) |
| Great Lakes St. Lawrence Seaway System Intermodal Map | Major ports in the St. Lawrence – Great Lakes Seaway | Canada/US | Ports | Great Lakes St. Lawrence Seaway System |
| Registre des entreprises | Cross-sectional data on active and inactive firms in Québec | Québec | Firms | Gouvernement du Québec |
Trade and transport datasets
| Dataset | Description | Geographic Coverage | Time Span | Frequency | Unit | Source |
|---|---|---|---|---|---|---|
| Air Carriers: T-100 Segment (US Carriers Only) | Time taken by planes between O/D pairs | Canada/US | 1990-Present | Monthly | Time | Bureau of Transportation Statistics |
| Border Crossing Entry Data | Number of trucks, trains and containers entering the US through border entry points with Canada | Canada/US | 1996-Present | Daily | Volume | Bureau of Transportation Statistics |
| Canadian Freight Analysis Framework | Data on freight flows by commodity and mode of transport within Canada | Canada | 2011-2017 | Annual | CAD, volume | Statistics Canada |
| Grain Supply Chain Dashboard | Movement of grain by rail in Canada at the station and corridor level | Canada | 2016-Present | Daily | Volume | Statistics Canada |
| Interstate Truck Trips by Origin and Destination | Provides the annual number of interstate trips undertaken by freight carrying trucks in the US | US | 2020-2022 | Annual | Trips | Bureau of Transportation Statistics |
| Spillover Simulator | Maritime capacity at risk of facing delays due to port disruptions affecting outgoing vessel movement | Global | 2022 (base year) | NA | USD, time | IMF PortWatch |
| Trade in Embodied CO2 (TeCO2) | Estimates on embodied carbon in final demand emitted anywhere in the world along global production chains | Global | 1995-2018 | Annual | metric tons of CO2 | OECD |
| Trade-and-Transport Dataset | Cost of transportation by mode and commodity between O/D pairs | Global | 2016-2021 | Annual | USD | UNCTAD, World Bank |
| TradeStats Express | Trade between US (at the state level) and the rest of the world | Global | 2009-2023 | Annual | USD | U.S. Census Bureau |