Article Preview
Top1. Introduction
Business intelligence has evolved from the earlier conception of decision support systems and executive information systems on the 80s and 90s originally put forward by Howard Dresner while still in the Gartner Group to the early days of the current agile, decentralized and data analytics driven orientation brought to focus by Thomas Davenport. Technological evolution and (consequent) data collection and processing capabilities have nowadays allowed to sustaining further traditional BI goals as well as moving toward the exploitation of new ones. In fact, besides the typical focus on historical data reconnaissance and future forecasting, we currently assist to some BI endeavours’ emphasis being given to inform present action, which is based on immediate analysis of high pace generated information. One may indeed position business intelligence endeavours’ outcomes as actionable at different organizational levels. However, formulating key performance indicators to inform managerial or strategic decisions will rely on different assumptions and requirements toward data, sources, scope, time horizon and pertinent analytical methods, than when targeting to inform immediate operational action (Kasemsap, 2016). Such holistic perspective on business intelligence brings two intertwined challenges to the existing frameworks: 1) what constitutes a fine level characterization of operational related information, which is actionable on 2) informing opportunities for immediate or short termed interventions.
In most industries, as those targeted by this work, operational level characterization encompasses the acknowledgement of the role of human operator. This work roots on the existing body of knowledge on the Human Factors (HF) field to contribute to furnish business intelligence projects’ frameworks. The proposed framework provides fine guidance on the consideration of dimensions surrounding operational context in the definition and analysis of coherent and articulated key performance indicators within the overall BI projects’ scope. Additionally, such aim is accomplished by complementarily extending the existing frameworks concepts (e.g. (Kimball & Ross, 2002)) particularly at the information modelling stage to account for human factors-related data as a significant dimension of analysis. Over the past 40 years, the HF research arena has combined numerous disciplines to study and apply physiological and psychological principles on engineering and design of products, tools, processes and systems toward improved adoption and user experience, reduced operative errors and increased safety and productivity (e.g. (Wickens, Lee, Liu, & Gordon-Becker, 2003)).
Furthermore, this work discusses a use case to depict the pertinence of the herein proposed framework application on BI project endeavours. The use case shows that it was possible to trace derived KPIs related to minor nonconformities on a production line to the operational level data on (work) process, operator and workplace. Both the solid theoretical grounding of the proposed framework and the discussion of the outcomes achieved through its usage on the reported use case unveil the contribution of this work as being one step forward on promoting the accountability of human factors on BI projects endeavours.
This paper is organized as follows: Section 2, Background, addresses business intelligence reference frameworks and introduces human factors’ topics considered in the scope of this work; Section 3 describes the methodological rational underlying the reported work. Section 4 delivers the proposed Business Intelligence for Human Factors - BI4HF framework. A use case of BI4HF is presented in Section 5 and Section 6 presents conclusions and points both limitations and future related research directions.