Discovering discovery: Data discovery best practices

Abstract

Data discovery is the art of going beyond answering specific questions. Given the goals of the organisation and a clean and reliable dataset, the ‘data detective’ is at his or her best in formulating intriguing questions. This paper looks at a variety of ways to stimulate the creative analytics process through the study of anomalies, the employment of segmentation, the avoidance of cognitive bias, and sidestepping the confusion between correlation and causation. With this set of tools in hand, the analyst must then communicate insights clearly in order to become a change agent via data.