By Catherine D’Ignazio and Lauren F. Klein (Cambridge, MA, USA: MIT Press, 2020, 314 pp.)
Reviewed by Daniene Byrne
Today, it seems like all our daily experiences begin or end with data, with our every move leveraged for data collection. We, more or less, accept this. We know that while our lives unfold, our habits, purchases, travels, and comments become an unceasing stream of descriptive bits collected and assessed elsewhere, in the cloud, slightly outside our consciousness, occasionally popping up to help out with location, shopping tips, or spellcheck. It is all an inescapable and intimate, yet also distant and intangible reality of contemporary life. We are awed at the astounding extent of data and the ease with which it all can guide us. We are inspired by data’s potential for problem-solving.
Yet data processes, like problems, are far from straightforward. Data are incomplete, their measurements are inaccurate, and sources can be flawed. From the most basic decision of what is or is not measured, to the how and why, and where that measurement happens, all data are not equally representative. When we substitute data as a reality proxy, our interpretation of truth and fact are all affected by which data are collected, the quality of those data, how they are analyzed and presented, and what or whom they are meant to represent. Data can only support a good decision when the data are thoroughly understood, flaws, and all. Sometimes, the features our data represent can be far from our reality. Additionally, opportunities for bias occur at all levels, from the specifics of data collection and compilation to its assessment and application. These incomplete representations may tell more about who has collected the data and why they care, than about individual lives and needs.
While our lives unfold, our habits, purchases, travels, and comments become an unceasing stream of descriptive bits collected and assessed elsewhere, in the cloud, slightly outside our consciousness, occasionally popping up to help out with location, shopping tips, or spellcheck.
In Data Feminism, authors Catherine D’Ignazio and Lauren F. Klein do not merely deal with data. They pair data with feminism. Here, feminism is deployed as a “shorthand for the diverse and wide-ranging projects that name and challenge sexism and other forces of oppression, as well as those which seek to create more just, equitable, and livable futures.” Data Feminism inspires readers to undertake a “journey toward justice and toward remaking our data-driven world.” Data Feminism means an approach to data and problem-solving that is both personal and political, one where data analysis and collection is an act that supports justice in decision-making. As D’Ignazio and Klein explain, “Data Feminism is a way of thinking about data, both their uses and limits, that is informed by direct experience, by a commitment to action, and by intersectional feminist thought.”
From one’s first encounter with Data Feminism, the political impacts of data analysis are central. The book’s jacket, designed by Ori Kometani, sets the tone. The bold font for the authors’ names and title harkens to the graphically charged work of Barbara Kruger, and floats on a background tiled with hundreds of tiny, yet legible protest posters declaring lines such as: “BLM; Dignity for All; Engage; Climate Rights are Women’s Rights; NOT PAID; We are the Majority; Be Kind; Yes, We Can!” and more. Each sign is a data point for a cause.
Within the text, D’Ignazio and Klein encourage readers to consider and engage the data’s potential. Recognizing, understanding, and participating in data processes with feminist wisdom is empowering, and surprisingly accessible. The mantra of second-wave feminism: “The personal is political,” is a central theme of the text. The book’s approach centers around Black feminist touchstones, including Kimberlé Crenshaw’s concept of intersectionality, Bell Hooks’s writing on justice, and the Combahee River Collective’s statement on feminist rights. While familiar to feminists, these concepts can be revelations to data analysts. Out of this movement, D’Ignazio and Klein derived seven core principles as their guiding structure and chapter themes: 1) examine power; 2) challenge power; 3) elevate emotion and embodiment; 4) rethink binaries and heuristics; 5) embrace pluralism; 6) consider content; and 7) make labor visible.
The Introduction describes the career of Christine Mann Darden. Darden was NASA’s brightest women computer (a term referring to the women at NASA who did all the mathematical calculations for spacecraft projections by hand; they were literally humans that computed). She was an African American woman, working at what was a segregated NASA office doing calculations for more senior staff. C. M. Darden collected and used workplace data to fight against discriminatory workplace practices which were keeping her from the cutting-edge research she was capable of. She eventually attained the high-level job she deserved only because she was able to use data to gain respect and draw attention to the issue of the unfairness of her treatment.
Data Feminism inspires readers to undertake a “journey toward justice and toward remaking our data-driven world.”
Chapter one demonstrates “Examining Power” is the first step toward change. Here, we learn the story of tennis champion Serena Williams, who connects her harrowing birth experience to the much greater, disturbing U.S. data on African American birth rates. She realizes the numbers reflect African American mothers’ lack of power. The second step “Challenge Power” occurs when data about disparities increase visibility for affected groups. “Elevation of Emotion and Embodiment” demonstrates there is value in knowing the emotional and physical context of data generation, as D’Ignazio and Klein value “people as living feeling bodies in the world.” “Rethink Binaries and Hierarchies” shows the danger in biased data sets arising from misguided assumptions, for instance, early versions of Facebook asked users to select “male” or “female,” excluding an option for nonbinary people.
This growing sense of data contexts, bias, and the emotional and ethical impacts of decisions lead to the next theme: “Embrace Pluralism”—a call to value communities’ work in preparing and sharing data that represent community members, with data understood and owned by them, instead of being collected and controlled by outsiders. “Consider Context” seeks to avoid misinterpretation, such as what occurred when multiple news reports on a single, horrific incident, the Boko Haram kidnapping, were interpreted as if they represented multiple incidents. “Making Labor Visible,” speaks to revealing massive hidden amounts of human labor that exist behind seemingly invisible technologies. Data Feminism concludes with a chapter on metrics and accountability. An outside audit describes how closely the researchers have practiced what they preach within Data Feminism. For instance, D’Ignazio and Klein make a point of extensively clarifying the additional labor that went into creating Data Feminism, citing many behind-the-scenes contributors.
Data Feminism is meant to celebrate community knowledge and to reach into communities and inspire them by offering multiple examples of projects and data collection that are powerful but not complex. The language is informal and much of the content features research from marginalized communities and individuals outside the academy. In the handbook style, the authors demonstrate how extraordinary and ordinary people have used data to illuminate injustice and inform policy, and how readers can too. The authors present themselves as woke, white, female, academics, with interdisciplinary backgrounds. They are aware of the limits of their own worldviews and grateful for their opportunities.
Examine power; challenge power; elevate emotion and embodiment; rethink binaries and heuristics; embrace pluralism; consider content; and make labor visible.
Mixing Theory, Example, and encouragement, Data Feminism is readily accessible to the interested lay reader, offering multiple jumping-off points for more information. It also encourages those who have broader data skills to engage its principles. The strong title grabs attention and celebrates the sources of its content, taking a risk of narrowing the audience. Throughout, multiple visual examples enrich the text, though some of the graphics were shrunk to fit within the hardcover format and I found myself reaching for the magnifying glass more than once. Data Feminism skillfully combines two distinct academic areas of expertise, feminist scholarship, and data science. Classified as digital culture, Data Feminism, thorough and timely, is a welcome addition to both fields. It is openly available as part of the MIT <strong ideas> series but is also worth owning and enjoying in hard copy.
Daniene Byrne recently earned her Ph.D. in technology, policy and innovation from the Department of Technology and Society within the College of Engineering and Applied Sciences, Stony Brook University, Stony Brook, NY, USA. Her dissertation, “Who Steers Automated Vehicle Policy?” analyzed stakeholder inputs and policy outcomes in guidance documents for automated vehicle safety. Email: email@example.com.