DATA SCIENCE APPLICATIONS IN GEO-INTELLIGENCE

Recent advances in offshore foundations installation and operation monitoring, data transfer and storage and computational resources have led to an acceleration of the use of data-driven methods in geotechnical applications. While geotechnical engineers have already started to adopt these methods, there are still a number of challenges to overcome to allow routine use of machine learning. To expose more engineers to data science techniques, a community-driven prediction event on the use of machine learning for pile driving predictions was held prior to the ISFOG2020 conference in Austin, TX. The prediction event showed the potential of machine learning techniques and how they can be combined with engineering knowledge and judgement to enhance predictive models. Machine learning can also be used in classification problems and a second example of soil type classification from CPT data is demonstrated. The different types of machine learning models and their advantages and drawbacks are discussed. Although there are only two examples presented in this paper, the potential scope of machine learning in geotechnical engineering is wider and the challenges for a more widespread use of data science techniques are outlined.