A Global Mining Dataset
3 days ago
- #mining
- #geospatial-analysis
- #data-processing
- The International Council on Mining and Metals (ICMM) released a Global Mining Dataset with over 8,000 mines and related assets worldwide.
- The dataset includes detailed information on asset types, commodities, and confidence factors.
- The author's workstation setup includes high-performance hardware (AMD Ryzen 9 9950X, 96 GB RAM, NVMe SSD) running Ubuntu on Windows 11 Pro.
- Prerequisites for analysis include Python 3.12.3, DuckDB v1.3.0 with extensions, and QGIS 3.44 for mapping.
- The dataset was cleaned and processed using DuckDB, producing a spatially-sorted Parquet file.
- Key findings include the most common primary and secondary commodity pairs (e.g., gold and silver).
- China, the United States, and Australia have the highest number of mining assets in the dataset.
- A heatmap of asset locations was generated, and specific analysis was done on Canada's Diavik Diamond Mine using OpenStreetMap data.
- Umbra's Synthetic Aperture Radar (SAR) satellite imagery was used to visualize changes at the Diavik Diamond Mine.