Information-enabled decision-making in big data scenarios

Data Science tools are enablers of effective Information-Enabled Decision-Making (IEDM). Assuming a sufficient quality of both the available information and the processes involved in information processing and decision-making (DM), IEDM is supposed to reduce the risk of wrong decisions as opposed to conclusions grounded on intuition or other questionable criteria. In this chapter, some fundamentals are first discussed, such as the relations between data and information and the distinction between fully structured DM and partially structured or unstructured DM. The concept of causal modeling is also introduced to emphasize the difference between correlation and causation, as relations between mathematical variables and empirical properties, respectively. Then, some methods to operationalize an IEDM problem are shortly presented. Definition and assessment of information quality is a relevant issue in IEDM, especially in big data contexts. The chapter defines information quality according to a semiotic approach , and a hierarchical structure composed by three layers, traditionally called “syntactic,” “semantic,” and “pragmatic,” is discussed. Finally, an example of IEDM applied for fault detection in production machines and for the identification of root causes is presented.