There is a silent crisis occurring in our industry – with few exceptions, data science teams are ineffective. The reasons for this are manifold, ranging from incorrect application of statistical theory to faulty philosophical underpinnings. This wide-ranging talk starts with practical, ‘bottom-up’ principles for doing effective data science before moving onto technical considerations. We interrogate fragile techniques before showcasing methods that will never fail to produce results.

Technical level: Introductory level/students (some technical knowledge needed)