This repository contains scripts for processing and harmonizing UN Comtrade Plus dataset spanning 1988 to 2023. The workflow transforms raw trade data into a clean, standardized PostgreSQL database with consistent product classifications and country codes.
- 35 years of international trade data (1988-2023)
- All data converted to the Harmonized System Rev. 2022 (HS22 or H6) standard from original reporter classifications (HS92, SITC2, etc.)
- Consistent ISO-3 Alpha codes for all countries
- Optimized PostgreSQL schema with proper indexing and relationships
- Extensive data validation and cleaning processes
- Product-level data at the finest granularity level (HS codes with six digits length)
- The full database is uses around 65 GB on disk
This represents a decade of experience working with international trade data, incorporating lessons learned from handling various data quality issues and inconsistencies inherent in cross-country trade reporting.
For researchers, analysts, and organizations requiring immediate access to this cleaned and harmonized dataset, a production-ready version is available for 995 CAD (~700 USD).
📊 Access the Complete Dataset 📊
What you get:
- Complete PostgreSQL dump with all processed data
- Comprehensive documentation and data dictionary
- Schema optimized for analytical queries
- Significant time savings (avoiding the 3+ day processing time)
This curated dataset represents substantial value-add over raw UN Comtrade data through systematic cleaning, harmonization, and quality assurance processes. The pricing reflects the extensive development time and expertise required to produce research-grade international trade data.
Note: This is a derived analytical dataset created through substantial processing of UN Comtrade data. Distribution complies with UN Comtrade terms of use and I cannot reshare the raw data.
Clean Comtrade database diagram:
Tidy derived database diagram:
