Abstract:

Nikki Stevens examines the shape of digital data in the context of software creation. Conversations about algorithmic injustice and biased datasets are meant to point us to the inequitable outcomes that software systems create. However, these conversations miss the data model – the structure that shapes algorithmic affordances and makes biased datasets possible. In this talk, Stevens describes the powerful impact data models have on every aspect of information storage and retrieval. They briefly review alternative methods for storing digital data, and offer a case study on the ways that data models reflect the ontological and epistemological frameworks of their creators. Ultimately, they argue that until we pay careful attention to data models, we will not be able to reckon with the social impacts of the data we collect.