Reimagining Data Science Methodology for Community Well-Being Through Intersectional Feminist Voices

The ethos surrounding data science as a sociotechnical phenomenon is multifaceted. The phenomenon embodies both advantageous and detrimental discourses. On the one hand, data science systems in healthcare offer novel technologies to help private and public institutions aid in better decision-making. On the other hand, facial recognition software often jeopardizes fundamental human rights with invasive and discriminatory algorithms. While making data science systems, practitioners are typically encouraged to execute project management methodology CRISP-DM (Cross Industry Standard Process for Data Mining) to complete projects successfully. Created for data mining projects, CRISP-DM guides the management of data science projects with six phases: business understanding, data understanding, data preparation, modeling, evaluation, and deployment. This work-in-progress conceptual paper uses an intersectional feminist framework to critically analyze CRISP-DM for data science projects. The reimagined intersectional CRISP-DM or InCRISP-DM methodology embraces iterative intersectional feminist interrogation to clarify six standard CRISP-DM workflow phases with four provocations: Learning & Praxis, Harm Reduction, Transformation and Accountability & Transparency. Future work appeals to bringing awareness of transnational risks that can emerge when applying western project management methodologies to countries of the Global South.