In this article, we introduce the term “data science intelligence” as the verified and validated qualitative and quantitative outcomes of the data science workflow. This framing marries the disciplines of science policy and data science in order to empirically ground a way forward for mitigating public value failures resulting from the implementation and use of data science algorithms and practices. After identifying the public value failures in the data science ecosystem, we discuss two public value failures which offer significant challenges and opportunities for data scientists and the organizations they serve. Finally, we pose the Participation, Access, Inclusion and Representation (PAIR) principles framework for organizations seeking to minimize the impacts of these failures via the creation of a taxonomy capable of deploying data science that reflects the values of the communities they aim to serve. Preliminary quantitative outcomes are shared while future work will engage its qualitative aspects.
Monroe-White, Thema and Marshall, Brandeis, "Data Science Intelligence: Mitigating Public Value Failures Using PAIR Principles" (2019). Proceedings of the 2019 Pre-ICIS SIGDSA Symposium. 4.