Process data analytics and machine learning have positively improved chemical manufacturing in terms of turning massive amount of data into actionable insights. From open source code sharing to on-line self-paced learning, there are a lot of resources for practitioners to learn new analytics methods. It is, however, more difficult to correctly apply these methods. The purpose of this invited session is to bring together process data analytics and machine learning experts from industry and academia to share their practical experience. Common misconceptions (for examples, more data means higher accuracy? correlation implies causation? analytics methods don’t require domain knowledge? ) will be addressed. Best practice (for examples, how to frame the analytics problem? how to select the right analytics method? How to avoid overfitting? How good is good enough?) and the importance of foundational concepts (such as statistics, data pre-processing, programmatic thinking) will be highlighted.

Bridging the Skills Gap in Chemical Engineering

Computational Molecular Science and Engineering Forum (21), Education (04), Applications of Data Science to Molecules and Materials (T3), Next-Gen Manufacturing (T6A)