Big Data, Recovery, & COVID-19: A Call for Black Feminist Data Analytics, Part III

Black people are dying from COVID-19 in alarming numbers. We know this because big data tells us so. This mortality data has been instrumental in calling attention to the ways systemic racism disproportionately results in higher death rates within Black communities.

Nonetheless, big data does not recover nor restore humanity to our mothers, fathers, sisters, brothers, children, and other family members and friends who are gone. Instead, univariate measures of death listed neatly in spreadsheets confound the messiness and chaos that death brings to people’s lives. Colorful, (almost cheerful) bar charts and interactive graphs and visualizations belie the gravity of what they represent. Thus, in as much as big data is valuable in identifying health disparities, it compresses Black humanity, a contested premise in the United States, into quantitative containers that offer a final dehumanizing insult in its inability to represent the Black lived experience. Indeed visualizing Black people’s death rates is more likely to cause, as bell hooks, puts it, “the conditions for even greater estrangement, alienation, isolation, and at time grave despair.”

The issues with big data are not new nor are they surprising. Critiques as well as suggestions for quantitative methods that can transform big data into humanistic visualizations that “rebalance unequal distributions of power” are helpful. Yet, these methods, even as useful as they might be, are unable to capture the full dimension of people’s lives, which is buried in big data.

What, then, is the remedy for big data on COVID-19 mortality? How does one recover Black humanity from mechanisms that are designed solely to count the dead? A variety of projects have sought to recover the lives of Black people lost to COVID-19 through stories. The Undefeated has developed a striking tribute, “Honoring Black lives lost to COVID-19" to memorialize the lives of Black people who have died. Each person’s life is lovingly depicted and honored with a small anecdote. The Haitian Times’s project, “Adieu” also focuses on highlighting the stories of members of the Haitian community living in the United States who have succumbed to COVID-19. Both projects reject the primacy of big data, relying on Black communities to help recover the lives of their family members with narratives. Yet, even as these projects offer opportunities to recognize Black humanity, they are bereft of stories about. the precarious nature of Black life in a society that has continued to rely on a racial hierarchy which made Black people vulnerable to contracting and dying from COVID-19. In this sense, despite their value, The Undefeated and Haitian Times’s projects are similarly reductive, to some degree, as big data in that they only reveal a fraction of the Black experience in relationship to death, disease, and data.

Black feminist data analytics also decenters big data and places an emphasis on the recovery of the Black lived experience through storytelling. However, it also critically interrogates the notion of recovery, probing, and questioning the significance of recovering the Black self from the dehumanizing ontological experience of living in the United States. bell hooks offers a useful frame for troubling the idea of recovery by asking:

hooks compels us to seriously consider what recovery means in the context of Black life and whether it is indeed possible to engage in a data science praxis that imbues humanity into big data.

Black feminist data analytics addresses the computational challenge of big mortality data by injecting noise and disorder back into the data analysis process. It rejects efforts to create structured ways of knowing and thinking, not only about death but also about the self. In practice, this means approaching recovery in a manner that mirrors the work of salvaging (retrieving) inaccessible, lost, corrupted, and damaged data in computing. It means starting from the premise that systemic anti-Black racism has corrupted and challenged attempts to achieve Black selfhood, the hard drive of Black life where memories and experiences are stored. Accordingly, the recovery of Black humanity in big mortality data on COVID-19 requires novel techniques and machine-learning approaches for mining large datasets of unstructured data (co-morbidities and chain of events) included in death certificates and mapping it onto corresponding data (medical racism, poverty, essential work, high-density living patterns) on structural racism. Doing this offers a praxis of recovery and data analysis that sheds deeper insight into the ways that racism disputes Black humanity.

Black feminist data analytics also require technologies of recovery that uncover information on Black life that has been encrypted or hidden as a protective measure from white supremacy. The full measure of Black joy and pleasure are often strategically tucked away and are rarely available for surveillance. Thus, Black feminist data analytics must selectively break this encryption of Black life with decryption keys or passwords, knowledge about Black life that regards the self as the embodiment of “collective reality past present, family and community.” From this vantage point, collecting and visualizing unstructured data from obituaries, death announcements, and family accounts to shed light on Black humanity must be done with community knowledge and consent.

Ultimately, the recovery of Black humanity from big data on mortality is a tricky affair. Wrestling with questions about what it is and for whom it serves creates fertile ground for Black feminist data analytics to conceptualize data science processes that offer a fuller and more humanistic view of Black life.

Look for the next essay in this series to learn more about Black feminist data analytics and the value they hold for analyzing COVID-19 racial data.


Kim Gallon is the Executive Director of COVID Black. Learn more about COVID Black by clicking here.

Redefining Health Data…Connecting Black Communities to their Data.