Data-centric knowing: A case study of oil exploration in the North Sea

Abstract

This thesis investigates the practice of using large amounts of sensory generated, uncertain data for making prospects of oil and gas reservoirs in the North Sea. The research is based on a 36-month case study of the exploration unit in a European-based, internationally oriented oil and gas company. The thesis is focused on data-centric knowing, that is, how knowing emerges across various technologies, data and practices, and how decisions need to be made on large amounts of instrumentally generated and uncertain data. The thesis has four goals: (1) analyse the practices of making plentiful and uncertain data form credible evidence in knowing practices; (2) show how knowing emerges across scales of time, place, practices and technologies, that is, in a knowledge infrastructure; (3) reflect on the research methods to address such knowing practices; and (4) consider the practical implications of data-centric knowing. The thesis aims to contribute to the field of information systems (IS). A theoretical framework is established using existing research that has addressed the mutual constitutive, emergent and necessarily incomplete relationship among technology, data and people in knowing practices. Empirical insight is provided into how data-centric knowing unfolds in subsea oil and gas exploration. The analysis outlines the practices necessary to navigate the uncertain situations involved in data-centric knowing, including how maintaining multiple forms of data and interpretations is generative to knowing, how practitioners deal pragmatically with surprises, the multiple orders of worth involved in the process and how credibility implies being accountable for how knowing emerges. The various tools and practices necessary to operate data-centric knowledge infrastructures are also addressed, including how data-handling work is generative to data analysis, how the work is collective and how knowing implies continuous data trajectories that require flexible organisation. On a methodological level, the analysis considers strategies for scaling studies of data-centric knowing, as well as the need to relate to theory development pragmatically.