Data-Driven Representation in the Era of Big Data

• In Event: Social Constructions of Data and the Implications of Data Representation

Abstract

Bureaucratic attachments to abstraction and synoptic visibility have long been leveraged in projects of statecraft. Critical studies of these historical data representations have pointed to the ways in which quantified, visualized phenomena—especially when leveraged for control—has a tendency to elide systemic social, historical, and political issues on the ground (Scott 1999, Murphy 2017). Presently, our algorithmic ability to capture and render social phenomena has expanded exponentially. Although these potent renderings likewise draw criticism and outright denouncement (Brayne 2018, Eubanks 2018, Noble 2018), optimism in the goodness of data-driven modes forges on in a range of sectors—a seeming panacea for the previously uninterpretable in everything from governance to activism. Enthusiasm surrounding data-driven logics is reinforced in by calls for transparency across sectors; ideas and decisions must be undergirded by data (or “data-driven”). Being data-driven here acts as a kind of accountability, drawing authority from its predecessor of “evidence-based.” New streams of data and new computational capability here become synonymous with evidence, and are further buttressed by private industry technologies and academic research. This cross-sector investment also acts as assurance of more data to come. Building from a two-year participant observation study with the LA city data team, alongside work with local anti-surveillance activists, this paper looks of this fraught but potent political economy of data-drivenness from the standpoint of urban governance—and specifically the day-to-day process of what it means to become data-driven — to ask what can(not) data, and algorithmic representations thereof, be made into evidence for?

Author

• Leah Horgan, University of California, Irvine