By Jane Rohrer
As a part of the University of Pittsburgh’s Year of Data and Society programming, Catherine D’Ignazio gave a talk on November 5 about her highly acclaimed book, Data Feminism (MIT Press 2019). The book, co-authored with Lauren Klein, offers new ways of thinking about data as informed by intersectional feminism. D’Ignazio began her talk by contextualizing and defining intersectional feminism: citing key contributions from Francis Harper, Kimberlé Crenshaw, and the Combahee River Collective, intersectional feminism examines how different factors of power, status, and bias--such as gender, race, and class—create overlapping systems of discrimination and inequity. D’Ignazio went on to state that, while intersectional feminism is concerned with gender and sexuality, “it is not only even about gender.” As she went on to explain, “it is actually about power—it’s about understanding and challenging who has it and who does not.”
Klein and D’Ignazio wrote Data Feminism to apply an intersectional feminist framework to data science. In her talk, she explained precisely why this is so important. Data science has historically been rooted in ideals of objectivity and detached neutrality. Because of this, data and the sciences—fields typically coded as masculine or male-dominated—have been, in the mainstream, conceptually divorced from the humanities. These latter fields have typically been coded as emotional, subjective, and thus feminine. By applying humanistic—indeed, data feminist—frameworks to data science, D’Ignazio and Klein argue that we can understand and dismantle the racist, classist, and sexist biases latent in data technologies we use every day, all over the world.
In her presentation, D’Ignazio pointed out that popular news media is full of examples of corporate and government actors funding and proliferating sexist, racist, and classist data products—hiring algorithms that demote women’s resumes, facial recognition software that either targets or altogether ignores women of color, and beyond. And, D’Ignazio argued, we should not see these as fluke occurrences; instead, intersectional feminism recognizes data and its collection as one of the many ways that existing power structures are maintained. Data feminism seeks to formalize and apply this intersectional method to data science itself.
As part of this formalization process, D’Ignazio outlined the “7 principles of Data Feminism:” “examine power, challenge power, rethink binaries and hierarchies, elevate emotion and embodiment, embrace pluralism, consider context, and make labor visible.” But beyond this theory, a central goal of the book Data Feminism was to provide readers with concrete, tangible examples of what data feminist approaches to data science might look like. Each chapter of the book outlines one of these seven principles, illustrating what it really looks like to examine or challenge power—how one might rethink binaries or elevate emotion, and why context is so important to the visibilzation of labor.
One example D’Ignazio offered in her talk was that of fellow Year of Data and Science presenter Mimi Ọnụọha's project, “The Library of Missing Datasets.” In it, Ọnụọha outlines “data sets that a reasonable person might expect to exist.” If data sets are, in part, thought to help us humans understand and solve issues of pressing social need, Onuoha critiques the stunning absence of such high-need databases such as trans people killed or injured in instances of hate crime, people excluded from public housing because of criminal records, or a master database detailing if/which Americans are registered to vote. Because data science is so extraordinarily funded and powerful, the glaring absences pointed out by Ọnụọha’s project check several of D’Ignazio and Klein’s seven principles: Ọnụọha begs us to examine power and power structures, and elevate our emotional reaction to these profound gaps in what data could be, or is interested in, telling us.
D'Ignazio’s talk powerfully pointed out that datasets are anything but “objective,” and there are, indeed, emotional aims behind many contemporary, common uses of data; as Data Feminism illustrates, these motives do the most damage when they’re not made clear by data scientists. Racist, sexist, classist data projects should illicit emotional responses. We can start by honoring emotional responses to data, which always represents an inherently emotional world.
To learn more about Catherine’s work check out her website. View her Year of Data and Society Presentation.