A High-Resolution Municipal Multi-Regional Input-Output Dataset for Korea, 2020

Overview This release provides a high-resolution municipal multi-regional input-output (MRIO) dataset for Korea for 2020. The dataset represents economic accounts across subregions within a consistent MRIO framework. Dataset The release contains the combined municipal MRIO table and separate files for domestic intermediate transactions (Z), technical coefficients (A), imports (M), final demand (Y, Y_d, and Y_m), value added (VA), and total output (X). Monetary values are reported in million KRW. Coverage The dataset covers 229 municipalities and 76 commodity sectors in Korea for 2020. Format The combined table is provided as municipal_mrio_korea_2020.parquet, while individual matrices are available in the matrices/ directory. The release also includes the municipality-sector index, dataset manifest, and diagnostic summary. Data Access The Parquet files can be read in any environment supporting the format. For convenient access in Python, the release includes korea_mrio.py. After extracting the release, users should preserve the directory structure and run Python from the extracted release directory. Individual matrices can then be accessed directly: import korea_mrio as mrio mrio.Z mrio.A mrio.M mrio.Y mrio.Y_d mrio.Y_m mrio.VA mrio.X Each matrix is loaded from its corresponding Parquet file when first requested. The loader also provides functions for calculating technical coefficients and the Leontief inverse from matrices of compatible dimensions: A = mrio.make_A(mrio.Z, mrio.X) L = mrio.make_L(A) Computing the full Leontief inverse requires substantial memory and processing capacity. Source Code The construction and diagnostic source code is provided separately at:https://github.com/eneplab/korea-municipal-mrio