Inductive Social Science Research as a Necessary Element of Data Science: Theory, Methods, and Scientific Convergence
College of Engineering
Organizational Performance and Workplace Learning
Dr. Don Winiecki
Data science is a new scientific field that seeks to interrogate large and heterogeneous data sets in order to aid understanding of currently intractable phenomena. It is often characterized as a type of artificial intelligence. However, data science research often actually requires participation from knowledgeable and experienced qualitative researchers. In this presentation we will describe the work of qualitative social scientists (Anthropology undergraduate and Sociology undergraduate research assistants and a Sociologist tenured in a College of Engineering) collaborating with data scientists who are attempting to develop and validate models and algorithms that reliably predict criminal activity from the `Panama Papers database` and which can subsequently be generalized to perform similarly with other heterogeneous data sets. Data science models focus on financial traces that may indicate criminal behavior in the ‘Panama Papers’. Here, we will report specifically on our development and use of online ethnographic methods of research with a particular focus on understanding the social networks and culture of individuals identified in the Panama Papers data. This dual ‘data science and ethnographic‘ focus shows promise in supporting refinement data science models and algorithms.
Kappelman, Katherine and Hay, Bryant, "Inductive Social Science Research as a Necessary Element of Data Science: Theory, Methods, and Scientific Convergence" (2019). 2019 Undergraduate Research and Scholarship Conference. 81.