Peter Polack: Diagram

As algorithms continue to come under fire for their epistemic mistakes, statistical biases, and lack of transparency, a different dimension of algorithm operativity is going unchecked: how algorithm logic is designed to make particular phenomena appear to human perception.

Diagram is a series of annotated, algorithmically generated graphics that demonstrate the capacity of algorithms to produce perceptible phenomena, with a particular focus on the function of these phenomena in law enforcement in the United States.

The project addresses one of the latest iterations of “information-driven policing” technologies, the crime investigation software Palantir Gotham, which uses simple algorithmic rules to generate network diagrams of criminal interactions. To generate these diagrams, Gotham ingests crime data from a number of data sources, represents relationships among the data elements as a network, and allows crime analysts to adjust the parameters that produce these relationships. As a provocation and a visual study, Diagram depicts Gotham and its aestheticization of surveillance as a “playing out” of algorithmic logic, where algorithms and interactive conventions are designed to facilitate the sensible emergence of crime narratives.

Diagram is not a game, but a self-playing simulation that highlights procedural and aesthetic resemblances between crime analysis software and video games that generate narratives about law enforcement, criminal justice, and incarceration. From prison data management to prison simulators, a key aspect of these narratives is how they operate to brand representations of criminality and crime investigation procedure. What is at stake here is not the veracity of these narratives, but the capacity of algorithms to generate them in a compelling manner, and how this capacity is achieved systematically through design.

To identify play and elements of game design in software like Gotham is not to trivialize the lived consequences of its determinations, nor to reductively cast information-driven crime analysis as the fabrication of fictions about criminal suspects. Instead, Diagram regards play as the simulated movement between data collection and its articulation in data analysis software. Put another way, it highlights the discrepancy between crime data and its procedural re-composition. While this discrepancy is typically critiqued as a strategic oversight concealed behind a “black box,” the challenge remains to examine algorithmic conditions of perceptibility and the design conventions that support their rhetorical effectiveness. It is here that critical computational thinking and art making is especially needed today.

​

Peter Polack

Peter Polack is a designer and PhD candidate at the Department of Information Studies at UCLA. His work and research address how technical systems are designed to inform our perception, and the role of art-making in illustrating what technology makes perceptible. This focus is informed by his background in game design and data visualization, and by his work with analyzing the social impacts of police surveillance systems.