Data accumulates from exploration. Geodata is becoming bigger and more complex. There is also an unprecedented demand for minerals, driven by modern consumption and geopolitics. Geoscientists are pressured to improve exploration and handle bigger data, but the tools reside in data science. This talent gap has led to a new discipline – geodata science, which combines data science with geoscience. Artificial intelligence is a subdomain of geodata science. Few geoscientists are aware of geodata science and fewer academies are training such talent. This course will provide you with a comprehensive understanding of machine learning in geoscience using the geodata science framework. The course offers a hierarchical (from philosophy to practice) and cohesive pedagogy, spanning theory to practice, data generation to model validation, which demystifies opaque topics (e.g., how algorithms work, the data-driven philosophy, what is necessary data, and how to build a workflow).

Top takeaways:

-What is geodata science and machine learning, and how are they related.
-How to construct a common mineral prospectivity workflow.
-Where to find relevant public geodata.
-What are additional topics or subdomains of knowledge that would be necessary for the participant to further develop their skills in geodata science.
-How to use Orange Data Mining to create code-free workflows.

The course will be focused on geodata science, providing a basis for idiosyncratic questions that are typical of geoscientists who are unfamiliar with data science concepts. The course will be one-of-a-kind in the way it teaches cohesive theory, alternating with practicals using large public datasets, reinforcing learning and providing live answers to questions. The hands-on segments consist of constructing workflows using the user-friendly Orange Data Mining program. The construction of workflows will be supported by the presenters and support members, providing hands-on support at the individual level (e.g., to debug). Finally, attendees will be using their own computers to construct machine learning workflows that can then be re-purposed and re-used for their own projects/needs.