Unsupervised AI Exposes Hidden Rock Layers in Ghana Offshore Basin

A new study from researchers analyzing Ghana's offshore Keta Basin demonstrates the power of unsupervised machine learning to classify electrofacies and characterize porosity in data-scarce environments. The workflow, applied to Well~C, processed six standard wireline logs over a targeted depth interval, successfully identifying distinct rock units without relying on core samples. This approach offers a scalable solution for reservoir characterization in frontier basins where traditional coring is costly or unavailable. The paper, available on arXiv (2604.27126), details the methodology and key findings.