Data Feminism

Feminism, at its very core, aims to dismantle systems of oppression; however, the identification of which systems are oppressive and what kinds of beings are harmed by them has been the subject of debate in feminist circles for more than a century. Across the many waves of feminist movements and throughout the halls of humanities and social sciences departments around the world, feminist thought and feminist practices are heavily contested and often come into conflict. From difference and discord arise efforts to make both feminism and the world a more inclusive and just place. Yet there remain strong tensions over how to define feminism, how to realize feminists’ demands, how to apply feminist theory to a wide variety of subject matter, and how to bridge the gaps between theory and practice to build a better world for all. In an ambitious attempt to resolve some of those tensions in the field of data science, Catherine D’Ignazio and Lauren F. Klein’s 2021 book, Data Feminism, poses seven principles and strategies that are worthy of examination for those of us who might not hold the title of “data scientist,” but work with data nonetheless. As geographers and geography departments engage in efforts to improve our approaches to justice and equity, the principles of data feminism can be effective tools to guide our discussions for how to integrate feminism into our research practices and pedagogy, as well for how we apply geography in the public sphere.

Data Feminism aspires to unite both feminist activism and feminist critical thought behind a way of thinking about data that forefronts the labor and everyday practices surrounding the collection, production, analysis, and presentation of data. The authors put forth a critique of how governments and corporations employ data and statistics to preserve an unequal status quo, while also illustrating how data “can be wielded back” and can be “part of the solution” to an unequal distribution of power and resources (p. 17). As a result, the principles of data feminism stand to embolden everyone who works with data to see the people and communities involved in their production—who makes data, who is represented in the data sets, who benefits from them, how the relationships between those who count and those who are counted reinforce existing power structures, and how, through a deeper understanding of those relationships and the contexts in which data are produced, we can use data to challenge normalized oppression.

One of the greatest assets of the book is how it makes critical feminist thought accessible to people without a background in social theory. In recent years there has been growing recognition of the need for the application of critical and feminist theory in the natural sciences and across the human–physical divide in geography (see Lave et al. 2014; Carey et al. 2016; Eichhorn, Baker, and Griffiths 2020). Additionally, feminist geographers have made significant contributions to cartography, geographic information systems (GIS), and quantitative geographic methods for more than two decades (see Kwan 2001; Harvey, Kwan, and Pavlovskaya 2005; Pavlovskaya and Martin 2007). Applying critical and feminist thought to natural and quantitative sciences creates space for theoretical innovation and empirical richness from interdisciplinary research, but also for making a more welcoming geographic community that celebrates difference and recognizes the advantage of having diverse perspectives in making better science and better scientists. There continues to be, however, understandable concern from those who have not yet had the training to do this work concerning how to “get it right,” how to do justice to larger social inequities, and how to not cause more undue harm in the process, despite good intentions. Data Feminism can be a guidebook to introducing discussions of power, oppression, coliberation, and intersectionality, and how these topics are indeed relevant to anyone working with data. As social scientists, as well as experts in geospatial technology and environmental systems, geographers are uniquely situated at the methodological and multidisciplinary intersections where these conversations need to be had, encouraged, and taken seriously.

The book is broken up into seven chapters (plus an introduction and conclusion), each of which introduces a new principle of data feminism and how to apply that principle to data science. The introduction invites us to think seriously about why feminism matters for data science and “how can we use data to remake the world” (p. 10). Chapter 1, “The Power Chapter,” asks us to examine power “not only to understand it, but also to be able to challenge and change it” (p. 47). A data feminist approach requires an awareness of the structures that make data sets and data science possible, that are ultimately reflected in our results, analysis, and the impacts of our work. In Chapter 2, D’Ignazio and Klein emphasize the importance of challenging existing power structures through the manner in which we collect, analyze, imagine, and teach data science. The authors argue that critical analysis of power relations implicated in our work should be the basis for reconfiguring research and curriculum design to develop a better understanding of how knowledge is produced and how more inclusive knowledge production makes for the production of more complete knowledge. Chapter 3 draws from Haraway’s (1988) concept of situated knowledge to complicate the objectivity of data and to argue that elevating emotion and embodiment in the presentation of data can help to communicate the relations through which data are produced as well as stories that cannot be expressed through quantitative data alone. Through a deeper consideration of the axiom “What Gets Counted Counts,” Chapter 4 asks us to rethink the binaries, hierarchies, and categories that structure our data sets as also fundamental to the structure of power. By questioning classification systems that reinforce social constructions of race and gender, we can bring awareness to the harms inflicted by the production of scientific knowledge as well as the potential for a science that challenges power to heal past and ongoing injustices.

Chapter 5 introduces us to the range of characters and labor that are essential to data work to highlight the messiness of even the tidiest data sets and the importance of embracing pluralism. D’Ignazio and Klein, along with many feminist scholars, contend that accounting for the positions from which we approach knowledge production, synthesizing multiple perspectives, and centering perspectives that have traditionally been excluded from knowledge-making practices helps us work toward a deeper, richer understanding of the world (p. 136). In Chapter 6, “The Numbers Don’t Speak for Themselves,” the authors argue that context, where data come from as well as the social, cultural, historical, institutional, and material conditions within which they were produced, is crucial to the maintenance of data both for empirical accuracy and for just application (p. 152). Without context, data lose their significance and run the risk of reinforcing social inequities. Finally, Chapter 7 demands that we make labor visible by showing our work. Although the work that goes into the production of data is largely invisible to most people, taking the steps to emphasize the range of tasks and relationships that are necessary in the production of data can help to clarify the significance of our results, their context, and the credit due to our interlocutors, participants, and collaborators. The conclusion, in turn, invites the application of the principles of data feminism to our own work and beyond, converting our students, colleagues, and readers into fellow data feminists.

Although most geographers do at least some work with data, many of us might not consider ourselves data scientists. Additionally, despite the normalization of women’s rights and the need for gender equity, there are somehow still many of us who don’t consider ourselves feminists. Nonetheless, Data Feminism has relevance for a range of scholars: from those of us who work with qualitative data to those who work with quantitative data and mixed methods; for those of us who engage with research on cultural, political, economic geography, and environment–society relations to those who focus on the physical environment; for those of us who explicitly do work on geographies of gender and race to those who don’t; for those of us who identify as women as well as those who identify as men, nonbinary, and gender-fluid, at the intersection of our race, ethnicity, class, national origin, and standing within or outside of academia. A desire to responsibly and effectively conduct research and communicate our findings through our publications, in our classrooms, and to broader society is one thing that we all have in common. A data feminist approach can help us get there.

As a historically White and male-dominated discipline, geography is notorious for reinscribing colonial and patriarchal mindsets into our data and curriculum. From environmentally determinist to outright orientalist social and natural histories, geographers have been responsible for a lot of the science that has been used to justify the construction of race, as well as science that has erased the contributions of Indigenous peoples and women to what we now consider the fundamentals of geographic knowledge. Field sites occupied by predominantly White men continue to translate into textbook covers depicting those same men towering over the natural landscape, in turn fortifying the understanding that geography is comfortable as a White male discipline even in the twenty-first century. Although the demographics of the field are changing, representation continues to be a real issue that affects the experiences of nondominant geographers in the field as well as in the classroom. Even as awareness of the inequities in the production of scientific and geographic knowledge is gaining more and more traction across the discipline, there continues to be inertia over what we are actually doing to encourage diversity among our ranks and to create a more just geography and world. The implications of this hesitancy trickle down from the lack of diversity among geography faculty to representation of women and people of color in classroom material and data analysis, which, as a result, limits the ability to recruit undergraduates into the major, adding to a feedback loop that reinforces both underrepresentation of minoritized groups and the amplification of biases when it comes to collecting and interpreting data.

Applying a data feminist lens to geography can help us to identify the power and privilege within the discipline and how it shows up in our data sets. The gradients of power in the discipline—who has it and who doesn’t, who does the work and who doesn’t, who is represented in the data and who isn’t, who benefits and who doesn’t—have real impacts on what types of data get collected, who gets published, and who gets hired; in the absence of reflection, this can lead to the reproduction of colonialism through our work. Paying attention to the systems of power that structure our data means we also need to reexamine our systems of classification, how they affect what we measure and how we measure it, and how they reinforce existing power structures. Hierarchies oriented around gender and racial binaries in human subject data have been the most obvious and egregious, but that the same mechanisms of categorization similarly construct what we consider to be the natural world is often given less attention.

Some of the most useful ideas from the book arrive when D’Ignazio and Klein attempt to put data feminism into practice. The authors model what data feminist scholarship can look like by reflecting on their collaboration, their experiences of the project, and their own positionalities and privileges as White, cis-female, tenured professors at research universities that allow them to do this work. In keeping with the data feminist ideals they propose, they also employ examples of counterdata and nontraditional perspectives to add layers and complexity to the theory and research conducted within the academy and official research institutions. They hold themselves accountable to the principles they put forth by devoting an entire section of the appendix to clarifying the values and metrics that guided their methodology and decision-making throughout the writing process, particularly in regard to how the examples and theorists they cite address structural issues like racism, patriarchy, ableism, and distance from research site and subject matter.

Their goals were ambitious. D’Ignazio and Klein aspired to have 75 percent of citations on feminist scholarship from people of color, 75 percent of examples of feminist data projects by people of color, 75 percent of all citations and examples by women and nonbinary people, 30 percent of projects from the Global South, and 50 percent of projects from outside the academy—in addition to highlighting transgender examples and theorists, the power of community support networks, nonvisual methods of presenting data, Indigenous knowledge and activism, and how such principles can be applied without expensive technology or formal training (pp. 218–19). Even though they fell short of many of these objectives, not only are their experiences and learnings useful to those of us who similarly aspire to “walk the walk” of putting feminist theory and ideals into our everyday practice and work, but the obstacles they encountered in pursuing accountability serve to justify why we need data feminism. The challenges of doing this work result from existing structural inequalities embedded within the production of knowledge that need to be at the forefront of the conversations we have with one another as we work toward greater equity across academia.

Part of this task must include applying justice, diversity, equity, and inclusion (JEDI) strategies in our training of future generations of geographers and broader publics that are capable of critical thinking and civic engagement. Data Feminism can be especially useful here as it helps to ground what might seem like abstract social theory in the basics of how data work in the world to facilitate an understanding of data as always produced by unequal social relations and coconstitutive of the interests of people in positions of power. Additionally, although teaching about the perpetuity of social injustices often leaves students with a sense of hopelessness, emphasizing how data ethics and the possibilities of counterdata to engage and empower rising scientists and communities in ways that challenge how data have been instrumentalized for oppression can move toward more hopeful and transformative experiences in the classroom. While pushing students to see the nuances of the impacts of data, we can elucidate the challenges of working with different data sets while maintaining skepticism for the positivist technocratic solutions that are echoed in popular media and textbook interpretations of data work. As a result, students can better understand that the problem is not just in terms of data or a lack of data, but in relation to the culture in which we create data.

A data feminist approach to pedagogy has the potential to integrate feminist principles into actionable and measurable learning objectives that can be applied across geography curricula. Although it might be more obvious for students to learn data feminist concepts and skills in human geography classes, curricula that target these objectives will also make better environmentally and technologically focused geographers and members of the public. Students can better develop skills of reflexivity, capacities for examining their positionality, and the ability to responsibly collect, analyze, present, and interpret the data that they will continue to encounter throughout their lives. By accounting for the contexts and communities that are integral to data production, students can learn how to interpret larger issues and power structures at play, as well as how to do community work and see themselves as part of a community. Integrating these learning objectives into geography classes can help develop students into researchers and citizens who understand the impacts their work has on the world around them and make decisions to combat inequality and injustice.

The implications of our work as geographers go far beyond research and teaching. Geography’s relevance to climate change, economic development, and geopolitics makes it an important discipline for tackling some of the greatest challenges of our time. If we want our findings and values as geographers to have an impact on and value to society, we need to work on making geographic knowledge publicly available and legible to a wider audience. This is often not rewarded in academic circles, but it is vital to student recruitment and the continued existence of the discipline in addition to the communities we work with, serve, and inhabit. Understanding that knowledge production and distribution are effects of present inequities that are in part maintained by who gets to make data and who gets to use data is essential for moving beyond said inequities and abolishing what Gilmore (1993) described as a knowledge apartheid. As academic and professional geographers, we benefit directly from the labors of communities we work with and communities that are affected by the issues we study. Biogeographers, arctic climatologists, and GIScientists are just as culpable as those of us who examine human–environment and sociocultural relations.

At the same time, human geographers are not alone in using methods like participatory action research in tandem with communities so that they can also use the data and results for their own activism and problem solving. Participatory GIS and participatory physical geography both have made entries into community-based research that can have radical social as well as environmental impacts, even as the everyday politics and power relations involved in participatory research must be continually examined and made subject to reflection (Elwood 2006; Whitman, Pain, and Milledge 2015). Perhaps the most important lesson from Data Feminism, though, is recognizing that coliberation, where our own liberation is deeply bound with the liberation of our interlocutors, participants, students, and everyone affected by our research and teaching, should be an objective of the discipline of geography, one that we work toward through care and consideration for the people and places with which we work, as well as for ourselves. Any attempt at eradicating patriarchy and White supremacy must come from a place of conscientiousness that is integrated into our everyday lives and the work we do at all levels. Additionally, we must also remember that community participation and relationships of care are not boxes to be checked off after a day, a book, an academic term, or a field season, but must be part of an ongoing struggle against oppressive systems of power.

Yet I find myself wondering if data really need to be at the heart of solutions to social inequities. As D’Ignazio and Klein aptly note, context matters. Throughout the book, they provide clear examples of where and how data can support activist work. From better counts of maternal mortality and femicide to the Environmental Justice Atlas and the Anti-Eviction Mapping Project in San Francisco, they argue that data can convey relationships and stories that are at the heart of the feminist cause: to undo and ameliorate systemic and everyday forms of oppression. In the world we live in, where we measure what matters most and where resources are allocated based on statistics and equations that predict productivity, efficiency, ratings, and return on investment, it is common sense, particularly among data scientists, that data can be instrumentalized to make sure resources are distributed where they are most needed and to institute policies that protect the most vulnerable communities.

Although data can be wielded for “good” or even “justice,” there are certainly contexts in which collecting more data is not an appropriate solution and, in fact, creates more problems. The authors also provide numerous examples of where data have caused serious harm (e.g., crime data used in predatory policing) as well as examples of where less, no, or illegible data might better protect vulnerable communities. This is to suggest that the notion that data can just as easily repair the harm that they have helped to cause is idealistic at best; does justice need to be yet another datafiable process? If so, what are the implications for social movements that cannot or refuse to be datafied?

To truly create a more just world, we must also engage seriously with other ways of knowing and being beyond what data can easily represent, recognize that just because we cannot measure something does not mean that it does not matter, and concede that we don’t have to measure something to care about it. I don’t think the authors would disagree with this sentiment, but the book’s data-centric approach has its limits for realizing a feminist agenda if data become the predominant lens through which we view the world instead of justice. Nonetheless, Data Feminism provides us with a foundation for a feminist data science, and perhaps a feminist geography, and encourages us to care for how we work with data and the people and places they implicate as a core value of our geographic practice.