top of page
Image by SJ Objio
Title: Predicting Iron Ore Demand in the Chinese Steel Market Using Artificial Intelligence
Organization

Vale - DTI - EY

Period

January 2022 - January 2025

Status

Completed

Est. Budget

USD 2.6M

Est. Business Benefit

USD 20.0M

Indicative Value-to-cost

7.7x

Keywords

Iron ore demand, Chinese steel market, AI forecasting, market intelligence, cloud analytics, vessel movement

Benefit basis

Estimated value from improved demand forecast accuracy, faster market intelligence cycles, reduced manual modelling effort, and better commercial/logistics decisions.

Business Challenges 
The marketing intelligence team relied on multiple Excel-based models maintained by different analysts. As forecast complexity increased, this created duplication, inconsistent data sources, manual consolidation effort, slow cycle times, and risk of inconsistent outputs. The business needed a governed, integrated, cloud-based forecasting product to predict iron ore demand in the Chinese steel market and support faster, more reliable commercial decision-making.
Expanded Technical Solution
The project developed an integrated AI forecasting platform for Chinese steel-market demand and iron ore demand signals. It consolidated fragmented Excel-based analyses into a governed data model and automated pipeline. The model incorporated steel production indicators, blast furnace utilisation, port inventory, vessel movements, import/export flows, macroeconomic indicators, construction and infrastructure demand proxies, price signals, customer data, and historical purchasing behaviour. Users accessed forecasts and drivers through an online dashboard.
Technical Work Packages/Methods
  • Built a cloud-based data pipeline to ingest internal commercial data, vessel movements, market data, steel production indicators, inventories, and macroeconomic signals.
  • Replaced manual Excel consolidation with governed data models, version-controlled assumptions, automated refresh, and source traceability.
  • Developed forecasting models using time-series, regression, gradient boosting, ensemble modelling, and scenario analysis.
  • Created a dashboard showing demand forecast, forecast drivers, uncertainty bands, regional/customer segmentation, and market-signal alerts.
  • Implemented model monitoring, back-testing, and workflow governance for marketing-intelligence users.
Key Deliverables
  • Integrated data foundation for Chinese steel and iron ore demand forecasting.
  • AI forecasting models and scenario-analysis capability.
  • Cloud-based dashboard and automated reporting workflow.
  • Governance approach for replacing fragmented Excel models with a controlled analytical product.
Results and Value Created
The completed project reduced manual model consolidation effort and improved consistency, speed, and traceability of market intelligence. It enabled users to connect market signals, vessel movement, demand indicators, and forecast outputs in one environment. The estimated benefit reflects improved commercial planning, better demand-sensing, reduced analyst time, and improved sales/logistics decisions.
bottom of page