Data Science for the Geosciences

Spatial Data Aggregation is defined in our book as the integration of heterogeneous spatial or spatio-temporal data sources, with the aim of predicting the occurrence of hazards and resources. Much of the transition to net zero carbon relies on changing from fossil fuels to materials for renewable energy and batteries. Mineral exploration is therefore key to achieve this goal. Readers will engage in an active mineral exploration for battery metals in Cape Smith, Canada. Readers will learn how data science can be an effective guide in geological field work. To achieve this, several spatial information sources will be used, such as remote sensing, geophysical, and geochemical data. These sources need to be aggregated to guide field geologists in locating areas of interest with the aim of collecting samples. In that context, we introduce Bayes’ rule and Bayesian reasoning about knowledge and information. We emphasize the counterintuitive results of Bayesian reasoning: rare events are very difficult to predict even with very accurate data. Next, we cover the alternative to Bayes’: logistic regression. We emphasize the advantage and disadvantage of these opposite approaches.