Data Feminism: Challenging Power Through Critical Quantitative Methods
An intersectional approach examining how power structures influence data practices, with focus on who counts, who gets counted, and who decides what counts—particularly in South Asian contexts.
Welcome to our exploration of data feminism—an intersectional approach to data science and measurement that challenges traditional power structures in how we collect, analyze, and interpret data.
Throughout this presentation, we'll examine justice-oriented frameworks with a geographical focus on India and South Asia, drawing on groundbreaking work from scholars like D'Ignazio and Klein, whose 2020 MIT Press publication has reshaped how we understand the politics of data.
Join us as we investigate who counts, who gets counted, and who decides what counts in our data systems—and how we might reimagine these systems to better serve justice and equity.
VR
by Varna Sri Raman
ImpactMojo Decks
www.impactmojo.in
Setting the Stage: Why Data Feminism?
Data systems perpetuate power imbalances by excluding marginalized groups and overlooking significant contributions like women's unpaid work. Critical interpretation is essential as data is never truly neutral.
Data Reflects Power
Far from neutral, data systems mirror and reinforce existing societal power structures, often making invisible those already marginalized
Systematic Exclusion
Approximately 62% of data systems exclude marginalized populations, creating significant blind spots in our understanding
Uncounted Labor
A staggering 66% of Indian women's work remains uncounted in official statistics, particularly domestic and care work
Contextual Interpretation
"The numbers don't speak for themselves"—they require critical analysis and contextual understanding
Data feminism critically examines power relationships in data systems, challenging the myth of neutrality while recognizing data's potential for both oppression and liberation.
Core Principle: Examining Power
2
Analyze Power Dynamics
Scrutinize how power operates through data systems
2
Question Data Authority
Challenge who collects, interprets, and benefits
Reveal False Neutrality
Expose the myth of "objective" or "value-neutral" data
Recognize Dual Potential
Acknowledge data as both oppression tool and liberation pathway
At its foundation, data feminism begins by examining how power operates in the world. This critical analysis focuses on the often-invisible structures that determine who collects data, who interprets it, and ultimately, who benefits from its collection and use.
By challenging the widely-held notion that data is inherently "objective" or "value-neutral," data feminism reveals how all data is situated within specific social and political contexts. This perspective helps us understand data's dual nature—as both a potential tool of oppression and a powerful resource for liberation and justice.
What is Data Feminism?
Data feminism combines data science with feminist perspectives to examine power structures in data practices, challenging neutrality assumptions while offering more equitable alternatives.
Integrated Framework
Data feminism merges rigorous data science methodologies with feminist theoretical perspectives, creating an approach that is both technically sound and socially conscious.
It employs "feminism as a tool and critical thinking method" to question assumptions embedded in data practices and reveals blind spots in traditional approaches.
Power Analysis
Central to data feminism is a systematic examination of which communities benefit from data collection and analysis versus those who are marginalized or harmed.
This approach goes beyond gender alone—it addresses all forms of power inequality, including those related to race, caste, class, ability, and geographical location.
Data feminism provides a comprehensive framework for critically examining data practices while offering constructive alternatives. It challenges us to recognize that data is never neutral but always shaped by the contexts and power structures in which it is created and used.
ImpactMojo Decks
www.impactmojo.in
The Seven Principles of Data Feminism
Data feminism offers a framework to analyze and challenge power structures in data practices, emphasizing diverse perspectives, contextual understanding, and recognition of all contributions.
Examine Power
Analyze how power operates in data systems
Challenge Power
Actively work to dismantle data hierarchies
Elevate Emotion
Value emotional responses and embodied knowledge
Rethink Binaries
Question rigid categorizations and hierarchies
5
5
Embrace Pluralism
Include diverse perspectives and approaches
Consider Context
Situate data within broader social realities
Make Labor Visible
Acknowledge all work that enables data systems
These seven principles, developed by D'Ignazio and Klein, provide a comprehensive framework for practicing data feminism. Each principle addresses a different aspect of how power operates in data systems and offers guidance for more equitable approaches.
ImpactMojo Decks
www.impactmojo.in
Historical Context: Data as Control
Colonial data collection in South Asia (1872-1947) served as a mechanism of social control, establishing classification systems for caste, religion, and gender that continue to influence modern data infrastructure.
Colonial Census (1872)
First comprehensive census under British rule established categories that would shape social identity for generations
2
Social Classification (1880s-1920s)
Colonial authorities refined categorization systems for caste, religion, and gender to facilitate administrative control
3
Statistical Methods (1920s-1940s)
Development of specialized statistical approaches designed specifically for managing colonial populations
Post-Independence (1947+)
Many colonial classification systems persisted in independent India's data infrastructure
Data collection in South Asia cannot be separated from its colonial history. The British colonial administration in India instituted systematic census practices from 1872-1947 that weren't merely counting exercises but mechanisms of control.
Through classification and categorization, colonial authorities solidified social hierarchies around caste, religion, and gender—often hardening previously fluid social boundaries. Many statistical methods developed during this period were explicitly designed for managing colonial populations, and their legacy persists in contemporary data systems throughout the region.
ImpactMojo Decks
www.impactmojo.in
Beyond the Gender Binary
Data systems worldwide enforce binary gender categorization, erasing diverse gender identities despite legal recognition in countries like India, leaving millions of transgender persons statistically invisible.
Binary Data Systems
73% of global datasets enforce strict male/female categorization
Legal Recognition
India's landmark recognition of third gender (2014)
Diverse Communities
Hijra, Aravani, Kinnar communities historically erased
Census Invisibility
4.88 million transgender persons underrepresented
One of the most pervasive ways data systems enforce normative power is through rigid gender categorization. Approximately 73% of global datasets enforce strict male/female categorization, despite the rich diversity of gender identities found worldwide and particularly in South Asian contexts.
Despite India's landmark 2014 legal recognition of third gender, data systems continue to marginalize Hijra, Aravani, and Kinnar communities. This erasure is particularly troubling considering India's estimated 4.88 million transgender persons whose experiences and needs remain largely invisible in official statistics, limiting access to services and recognition.
ImpactMojo Decks
www.impactmojo.in
Case Study: India's Missing Women Workers
Despite women's continued economic activity in India, official statistics show a 49% decline in female labor participation since 1991 due to flawed categorization and measurement methods that fail to capture women's diverse work patterns.
Statistical Disappearance
Female labor force participation in India has shown a puzzling 49% decline in official statistics since 1991, despite evidence of women's continued economic activity.
This statistical disappearance masks the reality of women's economic contributions.
Categorization Problems
Much of women's agricultural work is categorized merely as "helping family" rather than recognized as employment.
This categorization stems from gendered assumptions about what constitutes "real work" versus domestic duties.
Methodological Failures
Standard data collection schedules frequently miss seasonal and informal work patterns that characterize many women's livelihood strategies.
One-time surveys fail to capture the complex, fluctuating nature of women's economic participation.
This case study reveals how seemingly objective statistical methods can systematically erase women's economic contributions. The apparent decline in female labor force participation represents not an actual withdrawal from economic activity but a failure of measurement frameworks to capture women's diverse and complex work patterns.
ImpactMojo Decks
www.impactmojo.in
Urban Bias in South Asian Data
South Asian statistical frameworks exhibit urban bias despite the region's predominantly rural population, failing to adequately capture rural women's economic activities and internal household dynamics.
65%
Rural Population
Percentage of South Asia that is rural despite urban-centric measurements
43%
Data Gap
Rural women's activities missed by standard economic indicators
78%
Household Analysis
Surveys using household as unit, concealing internal power dynamics
National statistical frameworks across South Asia reveal a persistent urban bias, privileging urban economic patterns and lifestyles despite 65% of the region remaining rural. This bias manifests in how economic activity is defined, measured, and valued—often using urban standards that fail to capture rural realities.
Rural women's diverse livelihood strategies, which frequently combine subsistence agriculture, natural resource management, artisanal production, and seasonal wage labor, are particularly poorly captured. Additionally, using the household as the primary unit of analysis conceals internal power dynamics and resource distribution, further obscuring women's specific economic conditions.
ImpactMojo Decks
www.impactmojo.in
Invisibilized Labor: The Care Economy
Women's unpaid care work—representing 39% of India's GDP if valued—remains economically invisible despite being essential to social systems. Time-use surveys show women spend 352 minutes daily on this work compared to men's 51 minutes.
If properly valued, unpaid care work would constitute approximately 39% of India's GDP, yet it remains entirely absent from economic calculations. Time-use surveys reveal stark gender disparities, with women spending an average of 352 minutes daily on unpaid work compared to just 51 minutes for men.
This invisibility stems from persistent conceptual divisions between "productive" and "reproductive" work—a false dichotomy that devalues essential care labor. Recent feminist economics initiatives have developed methodologies to measure and value care work, challenging traditional economic frameworks that render this labor invisible despite its fundamental importance to social and economic systems.
ImpactMojo Decks
www.impactmojo.in
Theoretical Framework: Standpoint Theory
Standpoint theory recognizes that knowledge is shaped by social position, with marginalized perspectives offering unique insights that can lead to stronger objectivity by revealing hidden biases in research.
1
Situated Knowledge
All knowledge is socially situated and reflects particular positions, making knowledge inherently partial rather than universal
2
Marginalized Perspectives
Those on society's margins have unique insights about power dynamics that are invisible to those in privileged positions
3
Strong Objectivity
True objectivity requires starting research from marginalized lives to reveal hidden assumptions and biases
4
Challenging Neutrality
Rejects claims of a "view from nowhere" that conceals the specific perspectives shaping research
Standpoint theory, developed by feminist philosophers of science including Sandra Harding and Patricia Hill Collins, provides a crucial theoretical foundation for data feminism. It argues that knowledge is always produced from specific social locations, making all knowledge partial and situated.
This framework suggests that marginalized groups often have epistemic advantages in understanding certain aspects of society due to their unique positioning. Rather than undermining objectivity, starting research from marginalized perspectives can lead to what Harding calls "strong objectivity"—a more comprehensive understanding that reveals blind spots invisible from positions of privilege.
ImpactMojo Decks
www.impactmojo.in
Theoretical Framework: Situated Knowledge
Situated knowledge recognizes that all knowledge emerges from specific positions and contexts, rejecting the myth of detached objectivity while emphasizing that acknowledging these perspectives strengthens scientific inquiry.
Positioned Perspective
All knowledge reflects particular social positions, cultural contexts, and historical moments
The "God Trick"
Claims of seeing everything from nowhere—objective vision without position—are impossible and misleading
Data Provenance
Data always produced from somewhere and by someone, carrying their assumptions and priorities
Embracing Positionality
Acknowledging the specific position from which data is collected strengthens rather than weakens analysis
Donna Haraway's concept of situated knowledge builds upon standpoint theory by emphasizing that all scientific knowledge is produced from particular embodied positions. She critiques what she calls the "God trick"—the illusion of seeing everything from nowhere that characterizes claims of pure objectivity in traditional science.
Applied to data science, this framework reminds us that data is always collected, categorized, and analyzed by specific people in specific contexts. Rather than pretending these positions don't influence outcomes, a feminist approach acknowledges them explicitly, allowing for more honest and rigorous analysis that recognizes its own limitations and perspectives.
ImpactMojo Decks
www.impactmojo.in
Data Coloniality: Global Power Dynamics
Modern data practices mirror colonial relationships, with most AI research concentrated in wealthy regions while South Asian data is extracted for analysis elsewhere, creating value that rarely benefits the originating communities.
78%
AI Research Concentration
Percentage of AI researchers working in North America, Europe, or China
92%
Data Extraction
South Asian data frequently collected but analyzed elsewhere
3.2%
Local Benefit
Portion of data value returned to communities providing the data
Contemporary data practices often reproduce colonial relationships in digital form. With 78% of AI researchers concentrated in North America, Europe, or China, South Asian data is frequently extracted and analyzed elsewhere, creating value that rarely returns to the communities providing the data.
This data coloniality manifests in research projects where Global South populations serve as data sources while Global North institutions control analysis, publication, and resulting intellectual property. Challenging these dynamics requires new models of data sovereignty, local ownership, and reciprocal relationships that break extractive patterns and ensure equitable distribution of both burdens and benefits in data ecosystems.
ImpactMojo Decks
www.impactmojo.in
Digital Divide: Intersectional Exclusions
South Asia's digital divide extends beyond access to meaningful participation, with intersecting barriers of gender, caste, language, and design control determining who benefits from digital systems.
Gender Gap
33% gender gap in internet access across South Asia
Compounded Marginalization
Caste, religion, region compound digital exclusion
Language Barriers
95% of programming languages use English
Design Control
Who designs systems determines who benefits
The digital divide in South Asia manifests not just in access to technology but in who can meaningfully participate in digital spaces and data systems. The 33% gender gap in internet access represents just one dimension of exclusion that intersects with other factors like caste, religion, and geographic location.
Language presents a particularly significant barrier, with 95% of programming languages based on English and most interfaces designed for users with specific educational backgrounds. These barriers reinforce the principle that who designs data systems largely determines who can benefit from them, highlighting the need for diverse representation in technology development.
ImpactMojo Decks
www.impactmojo.in
Challenging Methods: Beyond Quantification
Data feminism advocates for balancing quantitative methods with qualitative approaches to capture the full complexity of social phenomena and center community perspectives.
Quantitative Limits
Recognize boundaries of what numbers alone can represent
2
2
Mixed Methods
Combine quantitative and qualitative approaches
Narrative Context
Situate statistics within lived experiences
Participatory Design
Include community voice in methodology development
Data feminism challenges the primacy of quantification in social research, recognizing the inherent limits of reducing complex social phenomena to numbers alone. While quantitative methods provide valuable insights, they can flatten lived experiences and obscure important contextual factors that shape those numbers.
Mixed methods approaches that combine quantitative data with qualitative context offer more comprehensive understanding. This might include ethnographic observations, oral histories, or participatory research methods that elevate community knowledge. These approaches don't reject quantification but rather enrich it with the depth and nuance necessary for meaningful interpretation.
ImpactMojo Decks
www.impactmojo.in
Case Study: COVID-19 Data in India
India's COVID-19 data revealed significant gaps in reporting, with millions of deaths uncounted and systematic exclusion of marginalized groups. Rural areas, women, and lower-caste communities were particularly underrepresented, highlighting how data infrastructure decisions determine whose experiences count.
Undercounted Deaths
Official COVID-19 death tallies missed an estimated 3.4-4.9 million excess deaths according to multiple independent analyses. This massive undercounting reflects both data infrastructure limitations and political factors influencing reporting.
Rural areas, where healthcare facilities and testing were limited, saw particularly significant undercounting of both cases and deaths.
Intersectional Blind Spots
Gender disparities in reported cases reflected not biological differences but social factors—women's limited mobility and access to testing facilities led to systematic undercounting of female cases.
Caste dimensions of COVID-19 impact remained largely uncollected despite evidence that marginalized communities faced disproportionate effects due to occupational exposure and healthcare access disparities.
The COVID-19 pandemic in India provides a stark illustration of how data collection methods can systematically exclude marginalized populations. Urban-centric testing infrastructure, combined with mobility restrictions that disproportionately affected women, created significant blind spots in official data.
Community health workers and civil society organizations developed alternative documentation methods to address these gaps, revealing disparities invisible in government statistics. This case demonstrates how data infrastructure decisions—where testing centers are located, what information is collected during registration, how deaths are classified—have profound consequences for whose experiences get counted.
ImpactMojo Decks
www.impactmojo.in
Data Violence: How Data Harms
Data systems can inflict material harm on vulnerable populations through exclusion, surveillance, discrimination, and social sorting, as exemplified by India's Aadhaar system and algorithmic welfare delivery.
Biometric Exclusions
India's Aadhaar biometric identification system led to approximately 27.5 million people having benefits withdrawn due to authentication failures, disproportionately affecting manual laborers, elderly people, and those in rural areas.
Surveillance Targeting
Data-driven surveillance systems disproportionately monitor marginalized communities, reinforcing existing patterns of over-policing and creating chilling effects on political participation and freedom of expression.
Algorithmic Discrimination
Automated decision systems in social service delivery reproduce and amplify existing biases, creating new forms of technological exclusion that operate alongside traditional discrimination.
Social Sorting
Datafication functions as a mechanism of social sorting that classifies populations into categories determining their access to resources and opportunities, often reinforcing existing hierarchies.
The concept of "data violence" refers to the material harms that data systems can inflict on vulnerable communities. These harms go beyond privacy concerns to include denial of essential services, reinforcement of discriminatory practices, and creation of new forms of exclusion through technological requirements.
In India, the Aadhaar biometric system exemplifies these dynamics, with authentication failures leading to widespread benefit denials. Similarly, algorithmic systems in welfare delivery often encode existing biases into seemingly neutral technical systems, perpetuating discrimination while obscuring responsibility through technical complexity.
ImpactMojo Decks
www.impactmojo.in
Counter-Data: Community Responses
Marginalized communities across South Asia have developed their own data collection methods to challenge official narratives and make their experiences visible through documentation, mapping, and participatory approaches.
Documenting Violence
Dalit rights organizations have developed systematic documentation methods to record caste-based violence cases that official statistics often minimize or miscategorize. These counter-data initiatives provide evidence for advocacy and accountability.
Mapping Ownership
Women farmers' collectives are mapping land ownership patterns to challenge gendered disparities in resource control. These mapping projects make visible women's relationships to land that official records often obscure.
Participatory GIS
Communities use participatory geographic information systems to document resources, hazards, and community assets in ways that challenge official spatial representations and center local knowledge and priorities.
In response to data systems that render their experiences invisible, marginalized communities across South Asia have developed counter-data practices—alternative methods of documentation, analysis, and visualization that challenge official narratives and make visible what mainstream data systems ignore.
ImpactMojo Decks
www.impactmojo.in
Feminist Cartography in South Asia
Feminist mapping approaches in South Asia challenge power structures by documenting women's spatial experiences, revealing patterns of violence, identifying safe/unsafe spaces, and supporting resource rights claims that official maps often ignore.
1
Maps as Power
Understanding cartography as historically embedded in colonial control versus emerging community mapping practices
2
Violence Geographies
Projects like "Missing Women" mapping violence hotspots to reveal patterns invisible in aggregated statistics
3
Safety Mapping
Participatory documentation of safe/unsafe urban spaces to inform policy and design interventions
4
Resource Mapping
Counter-mapping land rights and common resources to challenge official property records
Feminist cartography in South Asia challenges conventional mapping practices by centering the spatial experiences of marginalized communities, particularly women. Maps are never neutral representations but powerful tools that can either reinforce or challenge existing power structures.
Projects like participatory safety mapping enable women to document spaces they experience as safe or unsafe, producing knowledge that formal crime statistics often fail to capture. Similarly, resource mapping by women's collectives documents relationships to common resources like forests, water sources, and grazing lands that official property records typically render invisible, supporting claims to resource rights.
ImpactMojo Decks
www.impactmojo.in
Feminist data approaches value embodied knowledge as essential, recognizing human bodies as both sources of data and sites of experience. This perspective challenges the separation between "objective" data and subjective bodily experience, particularly resonating in South Asian contexts with their holistic healing traditions.
Embodied Knowledge: Beyond Abstract Data
Lived Experience as Data
Data feminism recognizes lived experience as a valid and essential form of knowledge that complements and contextualizes quantitative information. Embodied knowledge—what people know through their direct physical and emotional experiences—provides insights that abstract statistics alone cannot capture.
Bodies as Sites of Data
The human body itself functions as both a site of data collection and potential resistance. Technologies like menstrual tracking apps demonstrate this duality—they can provide valuable health information to users while simultaneously functioning as surveillance mechanisms that commodify intimate data.
Health data systems often privilege biomedical frameworks over holistic understandings of wellbeing that include social, spiritual, and environmental dimensions central to many South Asian healing traditions.
Feminist approaches to data emphasize the importance of embodied knowledge—recognizing that bodies themselves are sites where data is both produced and experienced. This perspective challenges the artificial separation between supposedly "objective" data and subjective bodily experience.
In South Asian contexts, where diverse healing traditions have long recognized the interconnection between bodily experience and broader social and environmental factors, this reunification of embodied and abstract knowledge has particular resonance.
ImpactMojo Decks
www.impactmojo.in
Who Counts? Enumerating the Marginalized
Statistical systems often fail to properly count marginalized communities like Adivasi peoples, people with disabilities, and religious minorities, reflecting both technical limitations and deeper questions about societal inclusion.
Representation in Statistics
Who is counted in official data systems
Categorization Power
Who defines and applies statistical categories
Measurement Authority
Who decides what metrics matter
Interpretation Control
Who analyzes and explains statistical patterns
Census and survey methodologies across South Asia continue to undercount particular populations, including Adivasi (indigenous) communities often located in remote areas or with mobile livelihood strategies that make them difficult to enumerate using standard methods.
Similarly, people with disabilities are frequently undercounted due to narrow definitions of disability that reflect medical rather than social models. Religious minorities often appear in data primarily as statistical objects rather than subjects with agency in defining their own categories and priorities. These patterns of exclusion reflect both technical limitations and deeper questions about who is considered a full member of the polity worthy of being counted.
ImpactMojo Decks
www.impactmojo.in
Questioning Categories: Caste in Data
Caste enumeration in India demonstrates how data categories become political battlegrounds, with historical limitations on comprehensive collection and ongoing debates about whether statistical visibility advances equity or reinforces social divisions.
1931
Last comprehensive caste census conducted under British rule
1951-2011
Independent India's censuses collect data only on Scheduled Castes and Tribes, not comprehensive caste data
2010s
Growing debates on necessity of comprehensive caste enumeration
Present
Dalit scholars challenging methodologies of caste quantification
Caste enumeration in India illustrates how data categories themselves become sites of political contestation. The last comprehensive caste census was conducted in 1931 under British rule, with independent India's censuses collecting data only on Scheduled Castes and Tribes rather than all caste groups.
Contemporary debates around caste enumeration reveal competing concerns—statistical visibility may support equity policies, but categorization can also reinforce social divisions. Dalit scholars have led critical discussions about how caste should be measured, challenging both the absence of caste data and problematic collection methodologies that reinforce hierarchical understandings.
Gender-Based Violence Data: Gaps and Silences
Most gender-based violence goes undocumented in official data, with significant gaps in how violence is defined and measured. Alternative, community-led approaches are emerging to address these limitations.
86%
Unreported Violence
Percentage of gender-based violence incidents that go unreported according to National Family Health Survey
24%
Official Record
Cases that reach police or administrative data systems
9%
Service Access
Survivors who access support services after incidents
Gender-based violence (GBV) statistics represent one of the most significant areas where official data fails to capture reality. According to the National Family Health Survey, approximately 86% of GBV incidents go unreported, meaning administrative data captures only a fraction of actual prevalence.
Measurement frameworks themselves determine what forms of violence "count"—economic coercion, psychological abuse, and structural violence often fall outside standard definitions. Community-led documentation initiatives have developed alternative methodologies that better capture the full spectrum of violence, creating evidence bases for advocacy while prioritizing survivor autonomy and dignity.
ImpactMojo Decks
www.impactmojo.in
Alternative Metrics: Beyond GDP
Alternative economic metrics incorporate unpaid care work, well-being, and sustainability considerations that traditional GDP fails to capture. These approaches, informed by feminist economics and regional perspectives, provide more inclusive measurements of economic value.
Feminist Economics Critique
Feminist economists have long questioned traditional economic indicators like GDP that exclude unpaid care work and environmental impacts. These critiques call for more comprehensive measurements that capture the full range of economic activities.
Time-Use Valuation
Time-use surveys that document and value unpaid work provide alternative economic measures that reveal women's significant contributions typically invisible in conventional accounts. These methods show how economic priorities shift when care work is properly valued.
Gender-Sensitive Indicators
South Asian economists have developed regionally-specific gender-sensitive indicators that better reflect local economic realities and priorities. These include measures of time poverty, access to commons, and informal safety networks.
Well-Being Frameworks
Alternative economic frameworks drawing from Buddhist economics and other South Asian traditions emphasize well-being, sufficiency, and ecological sustainability rather than unlimited growth. These approaches suggest different metrics centered on quality of life.
Moving beyond conventional economic indicators like GDP requires developing alternative metrics that better capture the full range of economic activities and their impacts on well-being. Feminist economics provides crucial frameworks for this reimagining, particularly in valuing previously uncounted work and considering economic outcomes beyond growth.
ImpactMojo Decks
www.impactmojo.in
Data Justice Framework
A holistic approach that evaluates data systems through the lenses of fairness, accessibility, consequences, and power balances to address structural inequalities in data practices.
1
1
Fairness
Equitable data collection and system design
2
2
Accessibility
Ensuring data and algorithms are transparent and understandable
3
3
Consequences
Assessing impacts of inclusion and representation
Power Balances
Addressing asymmetries in data ecosystems
The data justice framework offers a comprehensive approach to evaluating and redesigning data systems with fairness at their core. It examines both the technical dimensions of data practice and the broader social and political contexts in which data systems operate.
This framework goes beyond narrow conceptions of data ethics to address structural inequalities in who benefits from data systems and who bears their burdens. It recognizes that justice in data systems requires attention to representation in datasets, fairness in algorithms, transparency in processes, and equitable distribution of both risks and benefits.
ImpactMojo Decks
www.impactmojo.in
South Asian Feminist Data Projects
Feminist organizations across South Asia are developing innovative data projects for advocacy and empowerment, including resource mapping in Kerala, gender statistics in Pakistan, and labor documentation in Bangladesh.
Kerala Resource Mapping
Women's collectives in Kerala have developed comprehensive resource mapping projects that document access to water, forest products, and common lands. These maps combine digital tools with traditional knowledge to support resource rights claims.
Pakistan Gender Data Portal
Feminist technologists in Pakistan have created an accessible data portal that aggregates gender-disaggregated statistics across sectors, making information on gender disparities more accessible to advocates and policymakers.
Bangladesh Worker Documentation
Garment worker unions in Bangladesh have developed mobile documentation systems for tracking labor conditions, wage theft, and safety violations, creating evidence for advocacy that centers worker experiences.
Across South Asia, feminist organizations are developing innovative approaches to data collection and use that challenge conventional methods while creating valuable resources for advocacy and community empowerment.
ImpactMojo Decks
www.impactmojo.in
Participatory Methods: Co-Creating Data
Participatory methods redefine research by engaging communities as knowledge creators rather than subjects, emphasizing shared ownership and democratic processes while challenging traditional research hierarchies.
1
2
3
4
1
Shared Ownership
Community control of data processes and products
2
Community as Researchers
Shifting roles from research subjects to knowledge producers
3
Democratic Knowledge Production
Research as collectively determined practice
4
Challenging Hierarchies
Disrupting expert/subject divides in knowledge creation
Participatory methods transform the traditional research relationship by positioning community members as researchers rather than merely objects of study. This approach emphasizes shared ownership of both data processes and products, recognizing the expertise that comes from lived experience.
Participatory approaches treat knowledge production as a democratic practice rather than a technical exercise reserved for credentialed experts. By challenging hierarchies between researchers and "subjects," these methods can produce more relevant and accurate knowledge while building community capacity for continued data work.
ImpactMojo Decks
www.impactmojo.in
Rural Women's Labor: Beyond Official Statistics
Rural women's economic contributions remain largely invisible in official statistics due to measurement challenges, categorization issues, and failure to account for seasonal variations and multiple simultaneous activities.
Rural women's labor presents particular challenges for conventional economic measurement. Agricultural work is frequently undervalued and undercounted, with women's contributions often categorized as "helping family" rather than as productive work in its own right.
Standard point-in-time surveys fail to capture seasonal variations in women's work patterns, which may change dramatically throughout the agricultural cycle. Additionally, women often engage in multiple livelihood strategies simultaneously—combining farming, livestock management, forest product collection, and craft production in ways that defy neat categorization. Women's actual land ownership also remains largely invisible when data is collected at the household level.
ImpactMojo Decks
www.impactmojo.in
Urban Informal Sector: Statistical Shadows
India's informal sector employs 92% of the workforce yet remains largely invisible in official statistics, with women workers disproportionately affected in precarious, uncounted roles like domestic service.
92%
Informal Employment
Percentage of India's workforce in informal employment
81%
Women Workers
Percentage of women workers in informal, precarious employment
4.2M
Domestic Workers
Estimated domestic workers excluded from labor force surveys
Urban informal sectors represent another area where official statistics fail to capture economic reality. With 92% of India's workforce engaged in informal employment, standard labor measurements designed for formal, regular employment miss most economic activity.
Women are particularly concentrated in precarious, uncounted work sectors like domestic service, where an estimated 4.2 million workers remain excluded from labor force surveys. Standard counting methodologies also struggle with mobile populations like construction workers, street vendors, and waste pickers who may not have fixed workplaces or regular schedules.
ImpactMojo Decks
www.impactmojo.in
Intersectionality in Data Collection
Intersectionality examines how multiple identity factors combine to create unique experiences of discrimination. Traditional data collection often fails to capture these complex intersections, requiring both conceptual and methodological innovations to accurately represent overlapping marginalized identities.
Beyond Single-Axis Analysis
Intersectional approaches move beyond analyzing social categories like gender OR caste in isolation, recognizing that people experience these identities simultaneously. This perspective reveals patterns invisible when examining categories separately.
Traditional data collection often forces respondents to identify with single categories, obscuring the complex realities of those with overlapping marginalized identities.
Methodological Challenges
Implementing intersectional analysis requires both conceptual shifts and practical changes in data collection methods. Variables must be designed to capture multiple dimensions of identity and allow for disaggregation across multiple axes.
Many current data infrastructures prevent truly intersectional analysis due to sample size limitations, categorical constraints, or analysis frameworks that cannot accommodate complex interactions between variables.
Intersectionality—a concept originated by legal scholar Kimberlé Crenshaw—provides a crucial framework for understanding how multiple forms of discrimination combine to create unique experiences that cannot be understood by simply adding separate forms of discrimination together.
Applied to data collection, this approach reveals how patterns of inequality manifest differently at the intersections of identity categories. For example, the economic status of Dalit women cannot be understood by looking separately at gender disparities and caste disparities—their specific position at this intersection creates unique experiences invisible in non-intersectional analysis.
ImpactMojo Decks
www.impactmojo.in
Case Study: Manual Scavenging Data
Government data dramatically undercounts manual scavengers in India compared to NGOs and researchers, revealing methodological failures and political incentives that minimize the true scale of this caste-based practice.
The enumeration of manual scavenging—the caste-based practice of manually cleaning human excreta—illustrates dramatic disparities between official and community-collected data. Official government counts identify just 42,303 manual scavengers nationwide, while Safai Karmachari Andolan (a movement of sanitation workers) estimates over 180,000 people still engaged in this practice.
This discrepancy reflects both methodological failures and political incentives to minimize the problem's scale. Standard enumeration methods miss many workers due to stigma, inconsistent definitions, and the temporary or illegal nature of much manual scavenging work. Additionally, state governments face both legal obligations to eliminate the practice and practical challenges in doing so, creating incentives to undercount.
ImpactMojo Decks
www.impactmojo.in
Case Study: Maternal Mortality Measurement
India's maternal mortality data shows progress but conceals disparities due to urban measurement bias and systematic undercounting of deaths among marginalized populations, particularly Adivasi women.
1
2
3
4
1
Official Statistics
India's MMR: 103 per 100,000 live births (2019)
2
Urban Measurement Bias
Data collection concentrated in areas with medical facilities
3
Demographic Disparities
Adivasi women's deaths systematically undercounted
4
Alternative Documentation
Community health workers developing supplementary tracking
Maternal mortality measurement in India illustrates how data infrastructure design can systematically exclude certain populations. While India has made significant progress in reducing maternal mortality (103 deaths per 100,000 live births as of 2019), these statistics conceal substantial disparities.
Measurement systems concentrate in areas with medical facilities, creating an urban bias that misses many rural deaths. Deaths among Adivasi (indigenous) women are particularly undercounted due to both geographic remoteness and cultural barriers to engagement with formal healthcare systems. Community health workers have developed alternative documentation methods to capture these missing data points, revealing mortality rates significantly higher than official statistics in many marginalized communities.
ImpactMojo Decks
www.impactmojo.in
Religious Minorities in Data Systems
Data systems don't just count religious groups but actively shape how they're defined and understood. In South Asia, these classifications have significant political implications, especially for Muslim communities who face systemic disadvantages that data can either reveal or obscure.
Statistical Construction of Categories
Data systems don't merely count pre-existing religious groups—they actively participate in constructing categories of "majority" and "minority" through their classification systems. Census categories that seem descriptive actually shape how religious boundaries are understood and experienced.
The very concept of "minority" emerges partly through statistical operations that count, compare, and classify populations according to specific criteria.
Data-Driven Interventions
The landmark Sachar Committee Report (2006) used data to document systematic disadvantages facing Muslim communities in India, particularly in education, employment, and access to credit. This influential report demonstrated how data could make visible patterns of exclusion previously dismissed as anecdotal.
Faith-based organizations increasingly develop their own data collection initiatives to challenge misrepresentations and ensure their communities' needs and contributions are accurately documented.
Religious identity in data systems presents particular complexities in South Asia's diverse religious landscape. Statistical categories that seem neutral actually participate in constructing religious boundaries, often hardening fluid identities into fixed classifications with political implications.
Muslim communities in India face particular challenges with representation in data systems, with the Sachar Committee Report documenting how marginalization has been rendered invisible in many official statistics. Faith-based organizations increasingly engage in counter-data practices, developing alternative measurement systems that better reflect their communities' realities and needs.
ImpactMojo Decks
www.impactmojo.in
Data for Whom? User-Centered Design
User-centered data systems prioritize marginalized communities' needs through participatory design and accessible visualization, ensuring data serves those most affected by the issues it addresses.
Marginalized-Centered Design
Prioritizing the needs and experiences of traditionally excluded groups in the design process
Participatory Methodologies
Involving communities directly in designing data systems that affect them
Accessible Visualization
Creating data presentations that work for diverse literacy levels and abilities
Needs Assessment
Determining community priorities before collecting data
User-centered design principles applied to data systems require fundamentally rethinking who those systems serve. Rather than designing for technical experts or policymakers alone, feminist approaches center marginalized communities in the design process, asking what information would be most valuable to them and in what forms.
Participatory design methodologies involve potential users throughout the development process, from initial needs assessment through testing and refinement. Accessible data visualization principles ensure that insights are available to people with varying levels of data literacy, education, and ability. This approach recognizes that the ultimate measure of a data system's value is whether it serves the needs of those most affected by the issues it addresses.
ImpactMojo Decks
www.impactmojo.in
Algorithmic Bias: South Asian Examples
Algorithmic systems in South Asia frequently embed biases that disadvantage marginalized populations across facial recognition, welfare distribution, healthcare, and employment - reinforcing existing patterns of discrimination while appearing objective.
Facial Recognition Failures
Facial recognition technologies developed primarily on lighter-skinned populations show significantly higher error rates when applied to South Asian populations, particularly those with darker skin tones. These failures have real consequences when such systems are used for identity verification in service delivery.
Automated Welfare Systems
Algorithmic systems used in welfare delivery often reproduce and amplify existing discrimination patterns. In India, automated benefit distribution systems have shown biases against rural residents, people with disabilities, and others who don't fit expected interaction patterns.
Medical Algorithms
Disease prediction models trained primarily on data from privileged populations often perform poorly for marginalized groups with different risk factors and healthcare access patterns. This can lead to both over- and under-diagnosis of conditions depending on how algorithms are calibrated.
Employment Screening
AI recruiting tools trained on historical hiring data tend to penalize candidates from backgrounds underrepresented in those historical patterns, including women in technical fields. These systems can codify and amplify existing workplace inequalities while appearing objective.
Algorithmic systems increasingly mediate access to essential services across South Asia, but these systems frequently embed and amplify existing biases. Facial recognition technologies developed primarily on lighter-skinned populations show significantly higher error rates for South Asian users, particularly those with darker skin tones.
Similarly, automated welfare systems often reproduce discrimination against marginalized groups through seemingly neutral technical processes. These issues highlight how apparently objective technologies can function as mechanisms for extending historical patterns of exclusion when developed without critical attention to power and representation.
ImpactMojo Decks
www.impactmojo.in
Beyond Inclusion: Transforming Systems
Data feminism seeks to fundamentally transform biased systems rather than simply making them more inclusive, focusing on power redistribution and community empowerment.
1
1
Beyond Token Representation
Moving past simply adding marginalized groups to biased systems
2
2
Redesigning Infrastructures
Building data systems with justice as foundational principle
3
3
Power Redistribution
Shifting from representation to decision-making authority
4
4
Community Empowerment
Reclaiming data as tool for collective action
Data feminism ultimately aims not just to make existing systems more inclusive but to transform those systems fundamentally. Simply adding marginalized groups to biased systems often creates new forms of harm while giving the appearance of progress.
True transformation requires redesigning data infrastructures with justice as a core principle rather than an afterthought. This means shifting from questions of representation (who is included in datasets) to questions of power (who makes decisions about data collection, analysis, and use). It involves reclaiming data as a tool for community empowerment rather than extraction, surveillance, or control.
ImpactMojo Decks
www.impactmojo.in
Environmental Data Justice
Environmental data collection exhibits significant inequities, with gaps in climate monitoring for vulnerable regions, dismissal of Indigenous knowledge systems, and emerging community-led efforts to document local environmental hazards.
Climate Data Gaps
Climate monitoring infrastructure is disproportionately concentrated in wealthy regions, creating significant gaps in data about vulnerable areas most affected by climate change. These gaps make it difficult to develop appropriate adaptation strategies for the most impacted communities.
Indigenous Environmental Knowledge
Indigenous communities across the Himalayan region have sophisticated systems for observing and documenting environmental changes that often predate and complement satellite data. These knowledge systems are frequently dismissed as "anecdotal" rather than recognized as valid data.
Participatory Pollution Monitoring
Communities affected by industrial pollution are developing participatory monitoring projects using low-cost sensors to document environmental hazards in their neighborhoods, creating evidence bases for accountability and advocacy.
Environmental data justice addresses the inequities in who benefits from environmental data collection and who bears its burdens. Climate data collection remains concentrated in wealthy regions, creating significant blind spots in understanding impacts on vulnerable communities.
ImpactMojo Decks
www.impactmojo.in
Data Feminism: Challenging Power Through Critical Quantitative Methods
A free resource created for public good by Impact Mojo.
Questions? Contact us at [email protected]
Find more resources at impactmojo.in and the OpenStacks project at unknown link
ImpactMojo Decks
www.impactmojo.in
Critical Quantitative Methods in Practice
Data feminism approaches transform traditional research methods by redesigning surveys, analyzing data with awareness of power dynamics, integrating qualitative context, and prioritizing excluded perspectives.
1
Survey Design
Carefully considering language, categories, and timing
2
Power-Aware Analysis
Statistical methods that acknowledge structural inequalities
3
Mixed Methods
Combining quantitative data with qualitative context
4
Inclusive Research Design
Centering typically excluded perspectives
Implementing data feminism in practice requires concrete methodological approaches that challenge traditional quantitative methods while maintaining analytical rigor. Survey design must carefully consider language accessibility, category construction, and timing to avoid systematic exclusions.
Statistical analysis with power awareness acknowledges how structural inequalities shape data patterns rather than treating all differences as natural or inevitable. Mixed methods approaches combine quantitative measurements with qualitative context to provide deeper understanding. Throughout the research process, centering typically excluded perspectives means explicitly designing studies to capture experiences often missing from conventional research.
ImpactMojo Decks
www.impactmojo.in
Digital Rights: Data Protection and Privacy
Privacy protections often function as privilege, disproportionately benefiting advantaged groups while marginalized communities face greater surveillance. Individual consent models fail to address communal data ownership prevalent in many South Asian contexts.
Digital Divide in Practice
Access to privacy varies dramatically across different communities
Communal Data Decision-Making
Traditional collective approaches to information sharing
Differential Privacy Impacts
Privacy protections impact different groups differently based on their social positioning. Techniques like differential privacy that add noise to datasets can render already marginalized groups statistically invisible while protecting privileged groups.
Gender, caste, and class differences in digital literacy create significant disparities in who can meaningfully exercise privacy rights in digital spaces. Complex privacy settings and terms of service exclude many users from controlling their data.
Consent and Community Identity
Individual-focused data protection frameworks overlook the communal dimensions of data in many South Asian contexts. Information about traditional knowledge, cultural practices, or community resources may have collective rather than individual ownership.
For many marginalized communities, privacy functions as a luxury good unavailable to those who must constantly prove their identity to access basic services. Surveillance becomes a condition of citizenship for those most dependent on state support.
Digital rights frameworks must account for how privacy and data protection operate differently across social groups. Standard privacy protections often function as luxury goods available primarily to privileged populations, while marginalized communities face intensive data extraction and surveillance as conditions for accessing essential services.
Individual consent models also poorly fit contexts where data has communal dimensions, such as information about traditional knowledge or cultural practices. Feminist approaches to digital rights recognize these power differentials and work toward models that ensure all communities can exercise meaningful control over data about their lives.
ImpactMojo Decks
www.impactmojo.in
Health Data: Sites of Contestation
Health data represents contested terrain where different knowledge systems clash. Issues include tensions between surveillance and autonomy in reproductive health monitoring, culturally limited mental health diagnostics, and the exclusion of traditional healing practices from mainstream health metrics.
Health data represents a particularly significant site where competing knowledge systems and power relations become visible. Reproductive health monitoring illustrates tensions between surveillance and autonomy, with period tracking apps and maternal health programs collecting intimate data that can serve both supportive and controlling functions.
Mental health data frequently misses cultural context, with diagnostic categories developed in Western contexts poorly capturing diverse expressions of psychological distress. Traditional healing practices remain excluded from health metrics despite their significant role in many communities' well-being. These contestations reveal how health data always embeds particular understandings of bodies, illness, and healing rather than neutral scientific facts.
ImpactMojo Decks
www.impactmojo.in
Data feminism challenges conventional measurement by developing creative methodologies for quantifying traditionally "unmeasurable" aspects of life while acknowledging the partiality of all metrics and the values they embody.
Measuring the Unmeasurable?
Quantifying Care and Emotion
Data feminism grapples with representing aspects of life traditionally considered "unmeasurable"—care work, emotional labor, and community resilience. These elements are essential to social functioning but resist conventional quantification.
Discrimination Metrics
Measuring discrimination and exclusion presents methodological challenges, as these experiences are often subtle, cumulative, and contextual. Developing metrics that capture both structural patterns and lived experiences requires innovative approaches.
Beyond Binary Measurement
Moving beyond simplistic success/failure measurements toward nuanced understanding of partial progress, contextual achievements, and complex outcomes requires new frameworks for evaluation and assessment.
Value-Explicit Measurement
Rather than claiming false "objectivity," feminist metrics explicitly acknowledge the values they embody and the perspectives they prioritize. This transparency strengthens rather than weakens their analytical power.
A central challenge in data feminism involves representing aspects of life traditionally considered "unmeasurable"—care work, emotional labor, community resilience, and other qualitative dimensions essential to social functioning but resistant to conventional quantification.
Rather than abandoning measurement entirely, feminist approaches develop creative methodologies that acknowledge the partiality and situatedness of all metrics. These approaches move beyond binary success/failure measurements toward nuanced understanding of complex outcomes. Instead of claiming false "objectivity," feminist metrics explicitly acknowledge the values they embody—recognizing that transparency about values strengthens rather than weakens analytical rigor.
ImpactMojo Decks
www.impactmojo.in
Open Data: Promises and Limitations
While open data initiatives aim to democratize information access, technical barriers, language restrictions, and digital literacy requirements create unequal access that particularly affects marginalized communities.
94%
Digital Literacy Gap
Percentage of government datasets requiring advanced digital literacy to access and use
82%
Language Barrier
Open data portals in English only despite linguistic diversity
12%
Community Access
Proportion of marginalized communities with meaningful ability to use open data
Open data initiatives promise democratic access to information but often reproduce existing inequalities in practice. Approximately 94% of government datasets require advanced digital literacy to meaningfully access and use, creating significant barriers for many communities.
Language presents another major barrier, with most open data portals available only in English despite South Asia's rich linguistic diversity. These technical and linguistic challenges create power differentials in who can actually use supposedly "open" data, with marginalized communities often requiring intermediaries to access information ostensibly about their own lives and communities.
ImpactMojo Decks
www.impactmojo.in
Data Visualization: Feminist Approaches
Feminist data visualization challenges conventional design by embracing uncertainty, incorporating cultural relevance, and explicitly acknowledging power structures that shape data.
Representing Uncertainty
Feminist data visualization moves beyond the "clean" aesthetic that implies certainty to show messiness, gaps, and ambiguity. These approaches visually represent uncertainty and partiality rather than projecting false precision.
Culturally Relevant Forms
Participatory visualization approaches draw on local visual traditions and culturally relevant symbols rather than imposing Western data visualization conventions. These culturally grounded visualizations can make data more accessible and meaningful.
Contextualizing Power
Critical visualizations explicitly show the power structures and social contexts that shape data rather than presenting statistics as floating abstractions. These approaches connect numbers to the systems that produce them.
Feminist approaches to data visualization challenge conventional design principles that emphasize cleanliness, simplicity, and certainty. Instead, they develop visual strategies that acknowledge messiness, uncertainty, and partiality—making visible the limitations and contexts of the data being presented.
ImpactMojo Decks
www.impactmojo.in
Quantified Self: Embodied Data Practices
Feminist approaches to personal data tracking challenge normative assumptions, embrace diverse bodies, promote data ownership, and shift from individual to community-focused monitoring.
Feminist Self-Tracking
Reframing self-monitoring through feminist lens that challenges normative assumptions
2
Beyond Normative Bodies
Developing health monitoring approaches for diverse body types and abilities
Data Sovereignty
Reclaiming personal data from corporate platforms and surveillance systems
Collective Monitoring
Moving from individualistic to community-based tracking practices
The "quantified self" movement of personal data tracking takes on different dimensions when viewed through a feminist lens. Rather than focusing solely on individual optimization, feminist approaches to self-tracking examine how monitoring practices relate to broader social structures and norms.
These approaches develop health monitoring systems that work for diverse bodies beyond normative ideals, challenging the one-size-fits-all design of many commercial tracking tools. They also emphasize data sovereignty—giving individuals and communities control over their personal information rather than surrendering it to corporate platforms.
Perhaps most distinctively, feminist approaches often shift from individualistic to collective monitoring practices, using shared data to identify patterns affecting communities rather than isolating health as purely personal responsibility.
ImpactMojo Decks
www.impactmojo.in
Building Data Feminist Institutions
Creating just data organizations requires ethical policies, diverse leadership, accountability systems, and community governance structures that center affected populations.
Ethical Policies
Organizational guidelines ensuring justice-oriented data practices throughout project lifecycles
Diverse Leadership
Representation across multiple dimensions in data teams and decision-making positions
Accountability Mechanisms
Structured processes for addressing harms and ensuring responsibility
Community Governance
Advisory boards and participatory structures centering affected communities
Translating data feminist principles into institutional practice requires developing organizational structures and policies that embed justice at every level. Ethical data practices must be codified in organizational policies covering the entire project lifecycle, from conception through data collection, analysis, publication, and archiving.
Diversity in data teams and leadership ensures multiple perspectives inform decisions, but must go beyond token representation to include meaningful power-sharing. Accountability mechanisms provide structured processes for addressing harms when they occur, while community advisory boards and governance structures ensure those most affected by data systems have genuine influence over their development and use.
ImpactMojo Decks
www.impactmojo.in
Teaching Data Feminism
Data feminism education integrates critical perspectives with technical skills, connecting abstract concepts to social realities while ensuring regional relevance through contextual pedagogy.
Critical Curriculum Development
Educational programs teaching data science and statistics increasingly incorporate critical perspectives alongside technical skills. These approaches integrate ethical questions and power analysis throughout the curriculum rather than relegating them to a single ethics module.
Data literacy education expands beyond technical manipulation to include critical evaluation of data sources, methods, and implications—helping students understand both how to work with data and how to question it.
Contextual Pedagogy
Teaching methods that connect abstract concepts to concrete social realities help students understand the stakes and contexts of data work. Case studies centering South Asian contexts ensure relevance and applicability for regional students.
Interdisciplinary approaches bring together technical education with insights from social sciences and humanities, creating data practitioners who understand both the how and the why of their work.
Teaching data feminism requires reimagining both curriculum content and pedagogical approaches. Critical curriculum development integrates ethical considerations and power analysis throughout technical education rather than treating them as separate concerns.
Pedagogical approaches that connect abstract data concepts to concrete social realities help students understand the real-world implications of methodological choices. Using case studies centered in South Asian contexts ensures relevance while building students' capacity to apply critical frameworks to their own communities and experiences.
ImpactMojo Decks
www.impactmojo.in
Data Feminism for Policy Advocacy
Data feminism empowers policy advocacy by revealing hidden inequalities, analyzing budgets through intersectional lenses, and fostering partnerships between researchers and activists to drive meaningful change.
1
1
Critical Evidence
Using feminist data analysis for policy advocacy
2
2
Budget Analysis
Examining financial allocations through gender and caste lens
3
3
SDG Monitoring
Tracking development goals with disaggregated metrics
4
4
Coalition Building
Connecting researchers with activists for impact
Data feminism provides powerful tools for evidence-based policy advocacy by revealing patterns of inequality invisible in conventional analysis. Feminist approaches to budget analysis, for example, examine how financial allocations impact different gender and caste groups, often revealing disparities concealed in seemingly neutral budgeting.
Monitoring Sustainable Development Goal progress with intersectionally disaggregated data helps identify which populations are being left behind despite overall improvements. These approaches gain impact through coalition building between researchers with technical expertise and activists with organizing capacity, creating partnerships that can translate data insights into policy change.
ImpactMojo Decks
www.impactmojo.in
Corporate Data Practices: Feminist Critique
Feminist perspectives reveal exploitative corporate data practices and propose ethical alternatives focused on reciprocity, fair labor conditions, and algorithmic justice.
3
4
Extractive Relationships
Critiquing one-way data collection without reciprocal benefits
Surveillance Capitalism
Examining platform labor exploitation through monitoring
3
Algorithmic Bias
Identifying gender and racial bias in commercial systems
4
Alternative Models
Developing business approaches centered on data justice
Feminist analysis provides a crucial lens for examining corporate data practices, particularly the extractive relationships many companies establish with users and workers. These critiques identify how platforms extract value from user data while providing minimal returns to the communities generating that value.
Platform labor, increasingly common across South Asia, often involves intensive surveillance and algorithmic management that exacerbate power imbalances between workers and employers. Commercial algorithms frequently embed gender and racial biases that disadvantage already marginalized users. In response, feminist technologists are developing alternative business models centered on data justice principles, demonstrating that ethical data practices can coexist with sustainable operations.
ImpactMojo Decks
www.impactmojo.in
Feminist data governance approaches prioritize collective rights over individual privacy, recognizing community ownership and developing alternative infrastructures like data trusts and cooperatives.
Data Governance: Beyond Individual Privacy
Communal Data Ownership
Developing models of data governance that recognize collective rather than merely individual rights to information. These approaches acknowledge that some data inherently concerns groups rather than isolated individuals.
Indigenous Data Sovereignty
Supporting indigenous communities' rights to control data about their lands, cultures, and practices. These principles recognize that colonial data extraction has often accompanied resource extraction.
Feminist Data Infrastructures
Creating data trusts and cooperatives that enable communities to collectively govern their data and share in benefits. These structures provide alternatives to both corporate and state control of information.
Beyond Property Frameworks
Challenging the conceptualization of data as property to be owned, moving toward relational understanding of data as embedded in social contexts. These approaches recognize data's connection to lived experiences.
Data governance approaches rooted in feminist principles move beyond individual privacy frameworks to address collective dimensions of data rights and responsibilities. These approaches recognize that the current emphasis on individual consent and control fails to address many of the most significant data justice issues.
Models like data trusts and cooperatives provide structural alternatives that enable communities to collectively govern information about their lives and share in any resulting benefits. Indigenous data sovereignty principles particularly emphasize the right of communities to control information about their lands, cultures, and practices—recognizing that data extraction has often accompanied and facilitated resource extraction in colonial contexts.
ImpactMojo Decks
www.impactmojo.in
Decolonizing Data: Practical Steps
Decolonizing data requires acknowledging marginalized knowledge producers, embracing diverse knowledge frameworks, including multiple languages, and ensuring communities receive fair benefits from their data contributions.
1
Citation Politics
Citing marginalized knowledge producers and acknowledging intellectual debts
2
Epistemological Pluralism
Recognizing diverse knowledge frameworks beyond Western science
3
Linguistic Inclusion
Challenging English-centric data systems and interfaces
4
Value Return
Ensuring communities providing data receive proportional benefits
Decolonizing data practices requires concrete actions that challenge extractive research relationships and center marginalized knowledge. Citation politics—deliberately citing scholars from marginalized backgrounds and acknowledging intellectual debts—helps correct patterns of erasure in knowledge production.
Recognizing diverse epistemological frameworks means valuing knowledge systems beyond Western scientific traditions, including indigenous ecological knowledge and community documentation practices. Linguistic inclusion challenges the dominance of English in data systems, developing multilingual interfaces and methodologies. Perhaps most importantly, decolonial approaches ensure communities providing data receive proportional returns, whether as financial compensation, service improvements, or enhanced decision-making power.
ImpactMojo Decks
www.impactmojo.in
Ethical data collection requires justice-centered frameworks including culturally responsive consent, benefit sharing, contextual risk assessment, and sustained community engagement.
Ethical Frameworks for Data Collection
Culturally Responsive Consent
Developing informed consent processes that account for different cultural contexts, power dynamics, and communication norms. These approaches go beyond standardized forms to ensure meaningful understanding and voluntary participation.
Benefit Sharing Agreements
Creating formal agreements that specify how communities providing data will share in any resulting benefits, whether financial, educational, or in terms of enhanced services. These agreements establish reciprocal rather than extractive relationships.
Contextual Risk Assessment
Evaluating potential harms from data collection with particular attention to vulnerabilities created by intersecting marginalities. These assessments consider both immediate and long-term risks across different dimensions of community life.
Sustained Engagement
Maintaining relationships with communities beyond the data collection phase, ensuring ongoing communication, verification of findings, and accountability for outcomes. This engagement treats communities as partners rather than merely data sources.
Ethical data collection requires frameworks that center justice and reciprocity rather than merely compliance with minimum standards. Culturally responsive consent processes recognize that meaningful informed consent looks different across contexts and requires attention to power dynamics, language accessibility, and community decision-making structures.
Benefit sharing agreements formalize commitments to ensure communities providing data receive appropriate returns, while contextual risk assessments consider potential harms with particular attention to vulnerable populations. Throughout the process, sustained engagement with communities maintains accountability and builds relationships beyond extraction.
ImpactMojo Decks
www.impactmojo.in
Activist Data Science: Case Studies
Activists across South Asia are reclaiming data science tools to advance social justice causes, from environmental monitoring to documenting labor violations, while prioritizing community empowerment and ethical data practices.
Across South Asia, activists are using data science tools to advance justice causes while reimagining how these tools work. The Environmental Justice Atlas project maps conflicts between communities and extractive industries, documenting pollution impacts that official environmental monitoring often misses.
Women's organizations have developed gender-based violence tracking systems that prioritize survivor privacy and agency while creating aggregate data for advocacy. Land rights mapping projects use participatory GIS to document traditional usage patterns erased in official property records. Labor organizations create secure documentation systems for tracking violations in factories and informal workplaces. These initiatives demonstrate how data tools can be reclaimed for community empowerment rather than control.
ImpactMojo Decks
www.impactmojo.in
Quantum Feminism? Future Directions
Feminist frameworks offer critical perspectives on emerging technologies, challenging technological determinism while creating space for alternative computing paradigms based on diverse knowledge traditions.
Emerging Technologies Through Feminist Lens
As technological developments accelerate, feminist analysis provides crucial frameworks for examining quantum computing, artificial intelligence, and biotechnology before their impacts become entrenched. These analyses identify potential harms and alternative development paths early in technology lifecycles.
Challenging Technological Determinism
Feminist approaches question narratives that present technology development as inevitable and unidirectional, highlighting how social choices shape technical possibilities. These perspectives challenge deterministic views that hide human agency and responsibility in technological systems.
Drawing on diverse knowledge traditions opens possibilities for computing paradigms beyond current Western frameworks, potentially incorporating concepts from South Asian philosophical traditions, indigenous knowledge systems, and other marginalized epistemologies.
Looking toward the future, feminist analysis provides essential frameworks for examining emerging technologies before their impacts become entrenched. These approaches challenge technological determinism—the notion that technology develops along a single inevitable path—by highlighting how social choices shape technical possibilities.
Drawing on diverse knowledge traditions opens possibilities for alternative computing paradigms that might incorporate concepts from South Asian philosophical traditions, indigenous knowledge systems, and other epistemologies marginalized in current technological development. These speculative approaches imagine data futures beyond current constraints, asking not just how to make existing systems more inclusive but how technologies might develop differently if centered on justice from their inception.
ImpactMojo Decks
www.impactmojo.in
Practical Tools for Data Feminist Practice
Data feminism moves from theory to action through practical implementation tools: audit kits to identify biases, community review processes to ensure ethical oversight, and documentation frameworks to analyze power dynamics.
Audit Toolkits
Structured methodologies for examining existing datasets and algorithms for bias, exclusion, and harmful impacts. These toolkits provide practical steps for identifying problems in data systems before they cause harm.
Community Review Protocols
Formal processes for involving affected communities in reviewing and approving data projects before implementation. These protocols ensure that those most impacted by data collection have meaningful input into how it proceeds.
Documentation Templates
Standardized forms that prompt researchers to analyze power dimensions of their work, including who benefits, who bears risks, and how existing inequalities might be reproduced or challenged.
Translating data feminist principles into practice requires concrete tools that practitioners can implement in their daily work. Audit toolkits provide structured methodologies for examining existing datasets and algorithms for bias, exclusion, and harmful impacts—helping identify problems before they cause harm.
ImpactMojo Decks
www.impactmojo.in
Beyond Critique: Building Alternatives
Data feminists in South Asia are creating justice-centered infrastructures, repositories, and technologies with alternative funding models that prioritize community needs over corporate interests.
Feminist Data Infrastructures
Technical systems designed from the ground up with justice principles embedded in their architecture and governance
Community Data Repositories
Collectively owned archives preserving information vital to marginalized communities on their own terms
Open Source Feminist Tech
Collaborative development of software and hardware centered on justice and accessibility principles
Collective Funding Models
Alternative financing approaches supporting feminist data initiatives outside traditional funding structures
Moving beyond critique to construction, data feminists across South Asia are building alternative infrastructures that embody justice principles from the ground up. These initiatives include technical systems designed with equity embedded in their architecture and governance rather than added as an afterthought.
Community data repositories preserve information vital to marginalized communities on their own terms, while open source feminist technology projects develop software and hardware centered on justice principles. These efforts often rely on collective funding models that provide alternatives to both corporate investment and traditional philanthropy, allowing communities to direct resources toward their own data priorities.
ImpactMojo Decks
www.impactmojo.in
Measuring Success: Feminist Evaluation
Feminist evaluation redefines success metrics by prioritizing community values, engagement quality, relationship building, and learning from failures rather than focusing solely on technical outcomes.
Community Priorities
Evaluation metrics reflecting what communities themselves value
2
2
Process Indicators
Measuring quality of engagement beyond just outcomes
Relationship Building
Valuing sustained connections as core success metric
Failure Transparency
Honest reporting on challenges as learning opportunity
Feminist approaches to evaluation reimagine how we measure success in data projects, moving beyond narrow technical metrics to consider broader impacts and processes. These approaches center evaluation metrics that reflect community priorities rather than imposing external definitions of success.
Process indicators measure the quality of engagement throughout a project, not just final outcomes. Long-term relationship building is valued as a core success metric rather than treated as merely instrumental. Perhaps most distinctively, feminist evaluation embraces transparent failure reporting—recognizing that honest assessment of what hasn't worked creates valuable learning opportunities and builds trust with communities.
Who Would This Number Surprise — And Who Would It Silence?
Statistical measurements reflect power dynamics, privileging certain realities while erasing others through data collection choices and methodological decisions.
Statistical measurements often amplify certain realities while obscuring others. The numbers we choose to collect reflect power dynamics.
67%
Data Blind Spots
Percentage of global health metrics lacking disaggregated gender data
42%
Survey Exclusion
Rural households missed in digital-only government surveys
91%
Citation Bias
Academic papers citing Global North researchers over local experts
18%
Data Collection
Surveys including questions about unpaid domestic labor
Every statistical choice privileges some voices while erasing others. Quantification itself is a political act.
ImpactMojo Decks
www.impactmojo.in
Data Storytelling: Narrative Power
Feminist data storytelling transcends numbers by embedding statistics within meaningful narratives, emphasizing community strengths, and honoring diverse cultural traditions in data communication.
Contextual Statistics
Using stories as essential context for numbers, recognizing that statistics derive meaning from the narratives that frame them. This approach bridges quantitative precision with qualitative depth to create more complete understanding.
Asset-Based Narratives
Moving from deficit-focused reporting that emphasizes problems to asset-based approaches that highlight community strengths and resources. This reframing challenges narratives that portray marginalized communities primarily through lack.
Cultural Data Traditions
Incorporating visual and oral storytelling traditions from diverse cultures into data presentation, creating more accessible and resonant communications. These approaches recognize that data visualization has cultural dimensions.
Data storytelling recognizes that numbers alone rarely drive change—the narratives surrounding data shape how it is understood and used. Feminist approaches to data communication use stories as essential context for statistics, recognizing that numbers derive meaning from the narratives that frame them.
ImpactMojo Decks
www.impactmojo.in
Looking Forward: Data Feminist Research Agenda
A data feminist research agenda for South Asia requires methodological innovations for intersectional experiences, analysis of emerging technologies, and institutional changes in research funding and evaluation to support participatory approaches.
Future Research Priorities
A data feminist research agenda for South Asia must address several key priorities in coming years. Methodological innovations are needed to better capture intersectional experiences and community-level phenomena that current approaches miss. These might include new survey designs, mixed methods frameworks, and participatory research protocols.
As emerging technologies like AI reshape data systems, feminist analysis must stay ahead of these developments to ensure they don't reproduce or amplify existing inequalities.
Institutional Support Requirements
Advancing this agenda requires institutional changes in how research is funded, evaluated, and disseminated. Interdisciplinary collaborations between technical fields and social sciences are essential but face structural barriers in academia and research organizations.
Funding mechanisms must evolve to support participatory methods that take more time and resources than conventional approaches but produce more just and accurate results. Universities, research institutes, and funding bodies must recognize and reward community engagement as core scholarly work rather than peripheral service.
Looking forward, a data feminist research agenda for South Asia must address several key priorities in coming years. Methodological innovations are needed to better capture intersectional experiences and community-level phenomena that current approaches miss. Interdisciplinary collaborations between technical fields and social sciences are essential but face structural barriers in academia and research organizations.
Advancing this agenda requires institutional changes in how research is funded, evaluated, and disseminated. Funding mechanisms must evolve to support participatory methods that take more time and resources than conventional approaches but produce more just and accurate results.
ImpactMojo Decks
www.impactmojo.in
Personal Reflection: Where Do We Start?
Begin your data feminist journey through personal reflection, community building, and intentional action. This path includes auditing your own data practices, connecting with like-minded individuals, identifying opportunities for collective change, and celebrating incremental progress toward data justice.
Individual Data Practices Audit
Begin by examining your own relationship to data—how you collect, analyze, interpret, and use information in your work. Consider who benefits from your current approaches and who might be excluded or harmed. This personal inventory creates awareness that can guide more intentional choices.
Building Communities of Practice
Connect with others interested in data justice to share resources, strategies, and support. These communities can provide both practical guidance and emotional solidarity for challenging conventional practices. Even small groups meeting regularly can generate significant momentum for change.
Collective Action Possibilities
Identify opportunities for coordinated efforts to change data practices in your organization, field, or community. These might include advocating for policy changes, developing shared resources, or creating alternative data systems that better serve justice goals.
Small Steps Toward Data Justice
Recognize that transformation happens through cumulative small changes as well as dramatic shifts. Identify specific, achievable actions you can take immediately while working toward longer-term goals. Celebrating incremental progress sustains momentum for ongoing work.
Implementing data feminist principles can seem daunting given the scale of change needed in our data systems. Personal reflection offers a starting point accessible to everyone, regardless of their position or resources. By auditing our own data practices and relationships, we can identify specific areas where we have agency to make changes.
Building communities of practice provides crucial support for this work, connecting individuals with shared values to develop collective wisdom and resilience. While systemic change requires coordinated action, recognizing the value of small steps helps sustain commitment to this long-term project of transformation.
ImpactMojo Decks
www.impactmojo.in
Resources and Next Steps
This section provides essential resources for continuing your data feminism journey, including key readings from South Asian scholars, regional networks for community support, and practical implementation tools to transform data systems through collective action.
Key Readings
D'Ignazio & Klein's "Data Feminism" provides a foundational framework, while works by South Asian scholars like Payal Arora's "The Next Billion Users" and Noopur Raval's research on platform labor offer regional perspectives. These texts build theoretical understanding while providing practical examples.
Regional Networks
Connect with established networks like Feminist Internet Research Network, Internet Democracy Project, and Design Beku that bring together practitioners across South Asia. These communities offer workshops, resources, and mentorship opportunities for those looking to deepen their practice.
Implementation Tools
Practical resources like the Data Justice Lab's audit toolkit, Design Justice Network's principles, and Whose Knowledge? campaign's guides offer concrete steps for putting principles into practice. These tools can be adapted to specific contexts and needs.
As we conclude our exploration of data feminism, we invite you to continue the conversation beyond today. The resources we've shared provide starting points for deeper engagement, from theoretical foundations to practical implementation tools. South Asian data feminist networks offer communities of practice where you can connect with others engaged in this work.
Remember that transforming data systems is collective work that happens across multiple scales and timeframes. Whether you're a researcher, practitioner, student, or community member, you have a vital role to play in reimagining how data can serve justice and equity. We look forward to continuing this journey together.
ImpactMojo Decks
www.impactmojo.in
Feminist Critiques of Data Systems
Feminist perspectives challenge data systems by exposing power imbalances, centering marginalized viewpoints, and questioning assumptions embedded in data collection and interpretation.
- Standpoint Theory: Knowledge is shaped by social position. Marginalized perspectives offer unique insights into power structures.
- Power Analysis: Data systems often reinforce dominant hierarchies while claiming objectivity and neutrality.
- Right to Refuse: Communities can reject harmful data collection as an act of resistance.
- Biased Design Examples: Binary gender options exclude non-binary identities. "Household head" reinforces patriarchal assumptions.
- Survey Framing: Questions themselves contain assumptions that shape what counts as valid knowledge.
ImpactMojo Decks
www.impactmojo.in
South Asian Data Realities: Categorical Invisibility
Data systems in South Asia systematically exclude marginalized groups through categorical omissions, biometric failures, and misclassification of women's work.
Three critical gaps in South Asian data systems systematically erase marginalized experiences.
Caste Invisibility in DHS
Demographic Health Surveys often omit caste categories. This erasure masks health inequalities faced by Dalit and Adivasi communities.
Aadhaar-Based Exclusion
Biometric authentication failures disproportionately affect manual laborers and elderly people. Missing fingerprints mean missing benefits.
Time-Use Misclassification
Women's unpaid household work gets categorized as "non-productive." This renders invisible the economic value of care work.
These examples demonstrate how seemingly technical choices encode existing power structures into data systems.
ImpactMojo Decks
www.impactmojo.in
Beyond the Myth of Objectivity in Quantitative Research
Quantitative research is inherently shaped by researcher positionality, sampling choices, and methodological decisions. True rigor requires acknowledging these influences and embracing more inclusive approaches.
- Situated Quantification: Numbers aren't neutral. Researcher positionality shapes which questions we ask and which populations we count.
- Sampling Bias: Who gets included? Convenience sampling often excludes marginalized communities lacking digital access or formal documentation.
- Alternative Methods: Disaggregated indicators reveal hidden disparities. Participatory statistics center community knowledge.
- Reflexive Coding: Document assumptions in variable creation. Acknowledge power relations in analysis frameworks.
- Data Humility: Recognize the limits of what numbers can represent. Some experiences resist quantification.
ImpactMojo Decks
www.impactmojo.in
Counter-Data: Feminist Methodologies in Action
Feminist organizations are challenging traditional data collection by implementing community-based methodologies that document women's experiences, identify safety concerns, and recognize invisible labor.
Jagori Safety Audits
Feminist NGO mapping unsafe urban spaces through community walks. Women document missing streetlights, isolated areas, and transportation gaps.
DataKind Gender Index Critique
Technologists challenging mainstream gender metrics. Their audits reveal how conventional indices obscure lived experiences of marginalized women.
SEWA Time-Use Mapping
Self-Employed Women's Association documenting invisible labor hours. Their participatory methods validate women's economic contributions in informal sectors.
ImpactMojo Decks
www.impactmojo.in
Resources for Data Feminist Practice
Resources spanning theory, practice, education, and community to support feminist approaches to data collection and analysis.
Academic Foundations
D'Ignazio & Klein's "Data Feminism" provides essential theoretical frameworks. Pairs well with Tanu Priya Uteng's research on gendered mobility.
Practical Guidelines
IDRC's gender-data guidelines offer actionable research protocols. Perfect for researchers needing institutional implementation strategies.
Community Education
Nirantar's quantitative literacy resources center marginalized women's knowledge. Materials designed for grassroots statistical empowerment.
Digital Community
Join online forums like Data Feminism Network and Feminist Data Collective. Connect with practitioners applying these principles globally.
Community-Led Research
Local families contributing to participatory data collection processes in rural South Asia.
Intergenerational Learning
Knowledge transfer between generations strengthens community-based data practices.
Collaborative Analysis
Inclusive approaches ensure diverse perspectives inform data interpretation and application.
ImpactMojo Decks
www.impactmojo.in