Emancipatory Data Science: A Liberatory Framework for Mitigating Data Harms and Fostering Social Transformation

The cross-cutting and interdisciplinary nature of data work has created an opportunity to engage more students from diverse backgrounds in data science and has expanded pathways for entry for future data professionals. However, without greater representation of Black, Indigenous, and other marginalized people of color in data science, we risk reinforcing existing systems of differentiated power that oppress as opposed to empower these groups. In this paper, the term emancipatory data science is coined to highlight the unique contributions of individuals who use their expertise to mitigate data harms for minoritized, and marginalized populations and to suggest a way forward for the data science workforce and research community given our increasingly algorithmic society.