Understanding Risks of Data Science Failures through Sociotechnical Approach

Abstract

A long-standing question for CHI and CSCW community is to understand data science work and everyday practices. Based on six months of in-person and virtual ethnography with a private organization located in United States and India, this study explores how data science practitioners articulate and manage risks of data science project failure. Aligning with the plurality of risk articulations and power laden risk protocols, the research uses sociotechnical lens to explore the affordances between social actors and technology for risk management. The multiple affordances identified will inform strategies to help practitioners in everyday risk identification and management. The goal of this research is to offer empirical insights and actionable frameworks that will assist practitioners in navigating complex sociotechnical risk management landscapes.