Who makes maps and who gets mapped? Using a comparative reading of three maps, this case study introduces the idea that data may be useful, but they are not neutral. Rather, they represent the interests and goals of the groups and institutions that are doing the data collection. These interests and goals may be liberatory, discriminatory, or something in between. In all cases, we argue that an analysis of social inequality is essential to understanding the ethical impacts of data collection and use. To aid such analysis, we introduce a model of power out of sociology called the matrix of domination. This model helps us understand why collecting data is political, why not collecting data is also political, and what actions we can take to address unequal social relations using data science.
Keywords: redlining, social inequality and oppression, missing data, counterdata, matrix of domination, who questions
Author Disclosure(s): Portions of this case study are excerpted from chapters 1 and 2 of Catherine D’Ignazio and Laren F. Klein, Data Feminism (Cambridge, MA: MIT Press, 2020).
Learners will consider examples of how data science risks reinforcing existing oppression and inequality in society, as well as ways to mitigate that risk
Learners will compare three maps that exemplify mapping and the collection of data from different perspectives and in the service of distinct interests, goals, and incentives
Learners will practice analyzing distinct types of power relations and their impacts on data collection and map-making practices
Learners will consider how the act of collecting data is political, and how missing data—the act of not collecting data—is also political
Learners will start to ask who questions about data science: Who creates the map? Who is mapped? Who does the map benefit? Who does it harm?
In 1971, the Detroit Geographic Expedition and Institute (DGEI) released a provocative map, Where Commuters Run Over Black Children on the Pointes-Downtown Track. The map (figure 1) uses sharp black dots to illustrate the places in the community where the children were killed. On one single street corner, there were six Black children killed by white drivers over the course of six months. On the map, the dots blot out that entire block.
The people who lived along the deadly route had long recognized the magnitude of the problem, as well as its profound impact on the lives of their friends and neighbors. But gathering data in support of this truth turned out to be a major challenge. No one was keeping detailed records of these deaths, nor was anyone making even more basic information about what had happened publicly available. “We couldn’t get that information,” explains Gwendolyn Warren, the Detroit-based organizer who headed the unlikely collaboration: an alliance between Black young adults from the surrounding neighborhoods and a group led by white male academic geographers from nearby universities.1 Through the collaboration, the youth learned cutting-edge mapping techniques and, guided by Warren, leveraged their local knowledge in order to produce a series of comprehensive reports, covering topics such as the social and economic inequities among neighborhood children and proposals for new, more racially equitable school district boundaries.
Compare the DGEI map with another map of Detroit made thirty years earlier, the “Residential Security Map” (figure 2). Both maps use straightforward cartographic techniques: an aerial view, legends and keys, and shading. But the similarities end there. The maps differ in terms of visual style, of course. But more profound is how they diverge in terms of the worldviews of their makers and the communities they seek to support. The latter map was made by the Detroit Board of Commerce, which consisted of only white men, in collaboration with the Federal Home Loan Bank Board, which consisted mostly of white men. Far from emancipatory, this map was one of the earliest instances of the practice of redlining, a term used to describe how banks rated the risk of granting loans to potential homeowners on the basis of neighborhood demographics (specifically race and ethnicity), rather than individual creditworthiness.
Redlining gets its name because the practice first involved drawing literal red lines on a map. (Sometimes the areas were shaded red instead, as in the map in figure 2.) All of Detroit’s Black neighborhoods fall into red areas on this map because housing discrimination and other forms of structural oppression predated the practice.2 But denying home loans to the people who lived in these neighborhoods reinforced those existing inequalities and, as decades of research have shown, were directly responsible for making them worse.3
Early twentieth–century redlining maps had an aura very similar to the “big data” approaches of today. These high-tech, scalable “solutions” were deployed across the nation, and became one method among many that worked to ensure that wealth remained attached to the racial category of whiteness.4 At the same time that these maps were being made, the insurance industry, for example, was implementing similar data-driven methods for granting (or denying) policies to customers based on their demographic characteristics. Zoning laws that were explicitly based on race had already been declared unconstitutional; but within neighborhoods, so-called covenants were nearly as exclusionary and completely legal.5 The effect of these policies, bolstered by maps like the example in figure 2, exacerbated racial inequality and have led to long-lasting, intergenerational impacts.6
Who makes maps and who gets mapped? The redlining map is one that secures the power of its makers: the elite white Christian heterosexual men on the Detroit Board of Commerce, their families, and their communities. This particular redlining map is even called Residential Security Map. But the title reflects more than a desire to secure property values. Rather, it reveals a broader desire to protect and preserve home ownership as a method of accumulating wealth, and therefore status and power, that was available—preferentially and unfairly—to white people. In far too many cases, data-driven “solutions” are still deployed in similar ways: in support of the interests of the people and institutions in positions of power, whose worldviews and value systems differ vastly from those of the communities whose data the systems rely upon.7
The DGEI map, by contrast, challenges this unequal distribution of data and power. It does so in three key ways. First, in the face of missing data, DGEI compiled its own counterdata. Warren describes how she developed relationships with “political people in order to use them as a means of getting information from the police department in order to find out exactly what time, where, how and who killed [each] child.”8 Second, the DGEI map plotted the data they collected with the deliberate aim of quantifying structural oppression. They intentionally and explicitly focused on the problems of “death, hunger, pain, sorrow and frustration in children,” as they explain in the report.9 Finally, the DGEI map was made by young Black people who lived in the community, under the leadership of a Black woman who was an organizer in the community, with support provided by the academic geographers.10 The identities of these makers matter, their proximity to the subject matter matters, the terms of their collaboration matter, and the leadership of the project matters.11
For these reasons, the DGEI exemplifies one of the core principles of what we have termed “data feminism”: challenge power. Challenging power requires mobilizing data science to push back against existing and unequal power structures and to work toward more just and equitable futures. To challenge power means we must take action against an unjust status quo. In this case study, we discuss one starting point for challenging power: Collecting counterdata—in the face of missing data or institutional neglect—offers a powerful starting point.
What do we mean by power? We use the term power to describe the current configuration of structural privilege and structural oppression, in which some groups experience unearned advantages—because various systems have been designed by people like them and work for people like them—and other groups experience systematic disadvantages—because those same systems were not designed by them or with people like them in mind. These mechanisms are complicated, and there are “few pure victims and oppressors,” notes influential sociologist Patricia Hill Collins.
In her landmark text, Black Feminist Thought, first published in 1990, Collins proposes the concept of the matrix of domination to explain how systems of power are configured and experienced.12 It consists of four domains: the structural, the disciplinary, the hegemonic, and the interpersonal (figure 3). Her emphasis is on the intersection of gender and race, but she makes clear that other dimensions of identity (sexuality, geography, ability, etc.) also result in unjust oppression, or unearned privilege, that become apparent across the same four domains.
The structural domain is the arena of laws and policies, along with organizations and institutions that implement them. This domain organizes and codifies oppression. Take, for example, the history of voting rights in the United States. The US Constitution did not originally specify who was authorized to vote, so various states had different policies that reflected their local politics. Most had to do with owning property, which, conveniently, most women could not do. But with the passage of the Fourteenth Amendment in 1868, which granted the rights of US citizenship to those who had been enslaved, the nature of those rights—including voting—were required to be spelled out at the national level for the first time. More specifically, voting was defined as a right reserved for “male citizens.” This is a clear instance of codified oppression in the structural domain.
It would take until the passage of the Nineteenth Amendment in 1920 for many (but not all) women in the United States to be granted the right to vote.13 Even still, many state voting laws continued to include literacy tests, residency requirements, and other ways to indirectly exclude people who were not property-owning white men. These restrictions persist today, in the form of practices like dropping names from voter rolls, requiring photo IDs, and limits to early voting—the burdens of which are felt disproportionately by low-income people, people of color, and others who lack the time or resources to jump through these additional bureaucratic hoops.14 This is the disciplinary domain that Collins names: the domain that administers and manages oppression through bureaucracy and hierarchy, rather than through laws that explicitly encode inequality on the basis of someone’s identity.15
Neither of these domains would be possible without the hegemonic domain, which deals with the realm of culture, media, and ideas. Discriminatory policies and practices in voting can only be enacted in a world that already circulates oppressive ideas about, for example, who counts as a citizen in the first place. Consider an antisuffragist pamphlet from the 1910s that proclaims, “You do not need a ballot to clean out your sink spout.”16 Pamphlets like these, designed to be literally passed from hand to hand, reinforced preexisting societal views about the place of women in society. Today, we have animated GIFs instead of paper pamphlets, but the hegemonic function is the same: to consolidate and reinforce ideas about who is entitled to exercise power and who is not.
The final part of the matrix of domination is the interpersonal domain, which influences the everyday experience of individuals in the world. How would you feel if you were a woman who read that pamphlet, for example? Would it have more or less of an impact if a male family member gave it to you? Or, for a more recent example, how would you feel if you took time off from your hourly job to go cast your vote, only to discover when you got there that your name had been purged from the official voting roll or that there was a line so long that it would require that you miss half a day’s pay, or stand for hours in the cold, or ... the list could go on. These are examples of how it feels to know that systems of power are not on your side and, at times, are actively seeking to take away the small amount of power that you do possess.17
The matrix of domination works to uphold the undue privilege of dominant groups while unfairly oppressing other groups. This remains true even as women constitute a majority of the world’s population. What does this mean? With respect to gender, for example, men constitute the dominant group, while all other genders experience structural oppression, which is evident in any research that studies wage gaps, wealth gaps, political representation gaps, and more. Sexism is the term that names this form of oppression. In relation to race, white people constitute the dominant group (racism); in relation to class, wealthy and educated people constitute the dominant group (classism); and so on.18
Using the concept of the matrix of domination and the distinction between dominant and oppressed groups, we can begin to examine how power unfolds in and around data. This often means asking uncomfortable questions: who is doing the work of data science (and who is not)? Whose goals are prioritized in data science (and whose are not)? And who benefits from data science (and who is either overlooked or actively harmed)?19 These questions are uncomfortable because they unmask the inconvenient truth that there are groups of people who are disproportionately benefitting from data science, and there are groups of people who are disproportionately harmed. Asking these who questions allows us, as data scientists ourselves, to start to see how privilege is baked into our data practices and our data products.20
As we saw with the DGEI example, collecting counterdata can be a powerful strategy for exposing the differential harms of an unequal world, particularly when institutions run by dominant groups neglect to quantify or investigate such harms. Lacking comprehensive data about women who die in childbirth in the United States, for example, ProPublica decided to resort to crowdsourcing to learn the names of the estimated seven hundred to nine hundred US women who died in 2016.21 As of 2019, they’ve identified only 140. Or, for another example: in 1998, youth living in Roxbury—a neighborhood known as “the heart of Black culture in Boston”—were sick and tired of inhaling polluted air. They led a march demanding clean air and better data collection, which led to the creation of the AirBeat community monitoring project.22
Scholars have proposed various names for these instances of ground-up data collection, including counterdata or agonistic data collection, data activism, statactivism, and citizen science (when in the service of environmental justice).23 Whatever it’s called, it’s been going on for a long time. In 1895, civil rights activist and pioneering data journalist Ida B. Wells assembled a set of statistics on the epidemic of lynching that was sweeping the United States.24 She accompanied her data with a meticulous exposé of the fraudulent claims made by white people—typically, that a rape, theft, or assault of some kind had occurred (which it hadn’t in most cases) and that lynching was a justified response. Today, an organization named after Wells—the Ida B. Wells Society for Investigative Reporting—continues her mission by training up a new generation of journalists of color in the skills of data collection and analysis.
A counterdata initiative in the spirit of Wells is taking place just south of the US border, in Mexico, where a single woman is compiling a comprehensive data set on feminicide: gender-related killings of women and girls.25 María Salguero, who also goes by the name Princesa, has logged more than five thousand cases of feminicide since 2016.26 Her work provides the most accessible information on the subject for journalists, activists, and victims’ families seeking justice.
The issue of feminicide in Mexico rose to global visibility in the mid-2000s with widespread media coverage about the deaths of poor and working-class women in Ciudad Juárez. A border town, Juárez is the site of more than three hundred maquiladoras: factories that employ women to assemble goods and electronics, often for low wages and in substandard working conditions. Between 1993 and 2005, nearly four hundred of these women were murdered, with around a third of those murders exhibiting signs of exceptional brutality or sexual violence. Convictions were made in only three of those deaths. In response, a number of activist groups like Ni Una Más (Not One More) and Nuestras Hijas de Regreso a Casa (Our Daughters Back Home) were formed, largely motivated by mothers demanding justice for their daughters, often at great personal risk to themselves.27
These groups succeeded in gaining the attention of the Mexican government, which established a Special Commission on Feminicide. But despite the commission and the fourteen volumes of information about feminicide that it produced, and despite a 2009 ruling against the Mexican state by the Inter-American Human Rights Court, and despite a United Nations Symposium on Feminicide in 2012, and despite the fact that almost all Latin American and Caribbean countries have now passed laws defining femicide or feminicide—despite all of this, deaths in Juárez have continued to rise.28 In 2009 a report pointed out that one of the reasons that the issue had yet to be sufficiently addressed was the lack of data.29 Needless to say, the problem remains.
How might we explain the missing data around feminicide in relation to the four domains of power that constitute Collins’s matrix of domination? As is true in so many cases of data collected (or not) about women and other oppressed groups, the collection environment is compromised by imbalances of power.
The most grave and urgent manifestation of the matrix of domination is within the interpersonal domain, in which cis and trans women become the victims of violence and murder at the hands of men. Although law and policy (the structural domain) have recognized the crime of feminicide, no specific policies have been implemented to ensure adequate information collection, either by federal agencies or local authorities. Thus, the disciplinary domain, in which law and policy are enacted, is characterized by a deferral of responsibility, a failure to investigate, and victim blaming. This persists in a somewhat recursive fashion because there are no consequences imposed within the structural domain. For example, the Special Commission’s definition of feminicide as a “crime of the state” speaks volumes to how the government of Mexico is deeply complicit through inattention and indifference.30
Of course, this inaction would not have been tolerated without the assistance of the hegemonic domain—the realm of media and culture—which presents men as strong and women as subservient, men as public and women as private, trans people as deviating from “essential” norms, and nonbinary people as nonexistent altogether. Indeed, government agencies have used their public platforms to blame victims. Following the feminicide of twenty-two-year-old Mexican student Lesvy Osorio in 2017, researcher Maria Rodriguez-Dominguez documented how the Public Prosecutor’s Office of Mexico City shared on social media that the victim was an alcoholic and drug user who had been living out of wedlock with her boyfriend.31 This led to justified public backlash, and to the hashtag #SiMeMatan (If they kill me), which prompted sarcastic tweets such as “#SiMeMatan it’s because I liked to go out at night and drink a lot of beer.”32
It is into this data collection environment, characterized by extremely asymmetrical power relations, that María Salguero has inserted her feminicides map. Salguero manually plots a pin on the map for every feminicide that she collects through media reports or through crowdsourced contributions (figure 4, top). One of her goals is to “show that these victims [each] had a name and that they had a life,”33 and so Salguero logs as many details as she can about each death. These include name, age, relationship with the perpetrator, mode and place of death, and whether the victim was transgender, as well as the full content of the news report that served as the source. Figure 4 (bottom) shows a detailed view for a single report from an unidentified transfeminicide, including the date, time, location, and media article about the killing. It can take Salguero three to four hours a day to do this unpaid work. She takes occasional breaks to preserve her mental health, and she typically has a backlog of a month’s worth of feminicides to add to the map.
Although media reportage and crowdsourcing are imperfect ways of collecting data, this particular map, created and maintained by a single person, fills a vacuum created by her national government. The map has been used to help find missing women, and Salguero herself has testified before Mexico’s Congress about the scope of the problem.
Salguero is not affiliated with an activist group, but she makes her data available to activist groups for their efforts. Parents of victims have called her to give their thanks for making their daughters visible, and Salguero affirms this function as well: “This map seeks to make visible the sites where they are killing us, to find patterns, to bolster arguments about the problem, to georeference aid, to promote prevention and try to avoid feminicides.”
It is important to make clear that the example of missing data about feminicide in Mexico is not an isolated case, either in terms of subject matter or geographic location. Feminicide is a widespread problem that occurs globally in all countries that normalize violence against women, including the United States. And the phenomenon of missing data is a regular and expected outcome in all societies characterized by these unequal power relations, in which inequality and oppression are maintained through willful disregard, deferral of responsibility, and organized neglect for data and statistics about those members of oppressed groups who do not hold power. So too are examples of individuals and communities using strategies like Salguero’s to fill in the gaps left by these missing data sets—in the United States as around the world.34 If “quantification is representation,” as data journalist Jonathan Stray asserts, then collecting counterdata offers one way to hold those in power accountable.35 Collecting counterdata demonstrates how data science can be enlisted on behalf of individuals and communities that need more power on their side.
How are the three maps depicted in this case study similar? How are they different?
Analysis of power: Use the matrix of domination to analyze the practice of redlining. Here you might ask students to create a 2 x 2 table like the one in figure 3 and fill in their thoughts. What aspects of redlining fall into the structural domain? What aspects fall into the disciplinary domain? The hegemonic domain? The interpersonal domain? Which domain(s) would you place the redlining map into?
Peer share: Reflect on your own identity, background, and life experience. What aspects of the world do you have unique insight into because of your identity? What aspects of the world do you understand very little about because of your identity and life experience?
What who questions would you ask of a new data-driven technology to assess whether it is equitable across different social groups? For a specific example, you can ask students to reflect on facial recognition as it’s deployed in:
Public space surveillance to try to detect criminals
Installed in public housing by landlords to let tenants into their homes
Boarding an airplane to match your name to your face
Copyright © the Author(s) 2021