Transforming Data Into Added-Value Information: The Design Of Scientific Measurement Models Through The Lens Of Design Theory

Transforming data into added-value information is a recurrent issue in the context of "big data" phenomenon, as new sources of data become increasingly available. This paper proposes to offer a fresh look on how data and added-value information are linked through the design of specific models. This investigation is based on design theory, used as an analysis framework, and on a historical example in the Earth science field. It aims at unveiling the reasoning logic behind the design process of models combining data science and domain knowledge in specific ways, especially involving not only knowledge about the physical phenomena but also on the measuring instrument itself. More specifically, this paper shows how specific efforts on exploring the originality of the new instrument compared to existing ones can result in designing performant models to transform new sources of data into information. This also suggests several important competencies to be involved in the model-design process: (1) a detailed understanding of the limitations of existing models (2) the ability to explore both the originality of the new source of data compared to existing ones (3) the ability of leveraging independent data sources.