EXIOBASE 3 is a global, multi-regional input-output database, offering a detailed view of the world economy and its environmental impacts. It covers 49 regions and 200 sectors from 1995 to 2022. This dataset is ideal for analyzing trade, economic structures, and environmental footprints.
The data is stored in Parquet format and partitioned by year, format (ixi/pxp), and matrix type (Z, Y, F_satellite, F_impacts) to allow for efficient querying.
The following examples demonstrate how to effectively query the EXIOBASE 3 dataset using DuckDB.
For complex queries, such as time-series analysis or combining large matrices, direct queries may time out. A more robust approach is to break down the query into smaller parts and combine the results in your client application (e.g., Python with Pandas, R, etc.). The examples below follow this best practice.
To calculate the total exports to the United States, you should query the inter-industry (Z) and final demand (Y) matrices separately and then sum the results.
Step 1: Get Exports from Inter-Industry Transactions
Step 2: Get Exports from Final Demand
Combine the total_z and total_y results for each origin_region in your client to get the final ranking.
Similarly, for time-series analysis, query each matrix across all years and then join the results.
Step 1: Get Time-Series from Inter-Industry Transactions
Step 2: Get Time-Series from Final Demand
Join the results on the year column and sum z_value and y_value to get the total annual exports.
This query identifies the top 20 regions by total CO2 emissions, summing all CO2-related stressors from the F_satellite matrix.