Natural Language Processing: Detecting evidence in text reports and comparing with existing structured data

Over 5,000 USGS reports were analysed using Natural Language Processing (NLP) and Machine Learning to detect potential environments for Copper. Over 1.5Million detections were made. The results were coarsely spatialised by country and shown on the map above. The larger the pie-chart the greater the tone of uncertainty / speculation. These data were displayed in context to existing USGS spatial layers on known and estimated undiscovered copper resources. Given a lot more reports and more granular geo-locations, further patterns may emerge to potentially spark new questions or lines of investigation.