NZIPL Johns Hopkins SAIS · Net Zero Industrial Policy Lab

Atlas of the
Global Clean Industrial Base

A trade-based intelligence platform mapping the value chains of clean technologies worldwide — from raw material extraction to final product deployment. Phase II data platform, 2025–2026.

10
Technologies
200+
Countries
30 yrs
Coverage (1995–2024)
51+
Solar HS Codes
2,700+
ORBIS Firms
Batteries
Biofuel
Electrolyzers
Geothermal
Heat Pumps
Magnets
Nuclear
Solar
Transmission
Wind

Value chain stages: Extraction → Processing → Manufacturing → Final Product  ·  Roles: Raw Material · Processed Material · Product Component · Final Product · Process Equipment

Interactive 3-tab panels per country × clean technology: bilateral trade network, value chain timeline (2003–2024), and country competitiveness scatterplot. Nuclear panels use 2021 data (most recent available).

☀ Solar
⚛ Nuclear
🔋 Batteries
⚡ Transmission
💨 Wind

Interactive tree visualisations of the full HS code hierarchy for each clean technology — from raw materials upstream to final products, with value chain stage and role annotations.

UNCTAD-style Sankey flow diagrams tracing trade volume through each stage of the value chain (2022 data). Material-specific sankeys for critical minerals are also available.

By Technology
Critical Minerals
📦

BACI Bilateral Trade

CEPII

Harmonised bilateral trade flows at HS 6-digit level. Backbone of all trade analytics and value chain mapping.

1995–2024 ~200 countries HS 6-digit 6 revisions
📗

Green Dictionary

NZIPL Curated

Expert-curated mapping of HS codes to clean tech × value-chain stage × product role. Enables precise value chain filtering.

10 technologies 4 stages 5 roles HS92–HS22
🔗

WCO HS Concordance

World Customs Organization

Cross-revision HS code concordance enabling consistent time series across HS92→HS96→HS02→HS07→HS12→HS17→HS22.

6 revisions 3,499 edges Network crosswalk
🏭

ORBIS Firm Data

Bureau van Dijk

Firm-level location and NACE code data for clean tech manufacturers, linked to Green Dictionary via NACE→HS concordance.

~2,700 firms Global NACE 4-digit
🤖

Predicted Competitiveness (PC)

NZIPL ML Model

Random Forest–derived competitiveness scores with SHAP feature importance. Identifies which trade patterns predict future industrial leadership.

155 countries 10 technologies 2003–2024 ranger + SHAP
🌍

World Bank Indicators

World Bank Open Data

GDP (current USD) and population series for normalising trade exposure (exports % GDP). Used across all scatter and ranking visualisations.

200+ countries GDP + Population 1995–2024