The previous entries, which are elements of a broader research project, explored Pittsburgh as one site where the confluence of technical expertise in computation, public policy, and architectural research helped configure present day understandings of the city as software. As we saw, notions of design and of the city were renegotiated in computational terms, and new architectural identities emerged that sought to disrupt the discipline through quantitative and computational logics.

Where does this history leave us? Can we imagine urban technologies in ways that refuse, or at least critically acknowledge, the complicated legacies of these systems in military-academic industrialism and managerialism? Can new urban technologies be designed with present day computational methods such as machine learning, computer-vision, and sensing in ways that eschew commitments to surveillance-capitalist logics? And, I repeat, what can architectural modes of inquiry do to illuminate, challenge, or subvert these logics?

The following paragraphs briefly describe a few recent projects that explore these questions through the design of inquisitive urban technologies. They are not meant to offer conclusive answers. Instead, they each help articulate an important question concerning urban platforms.

Tracing urban life

Can digital platforms for urban analysis challenge revenue and police-centred applications and engender engaged data publics able to consciously participate and critically intervene in evolving portraits of urban life?

The WYSIWYG project combines spatial analysis methods with recent developments in data science, machine learning, and computer vision to understand how urban spaces give structure to human activity. It reimagines computationally William Whyte's study of 'The Social Life of Small Urban Spaces', which used film, qualitative observations, and clever counting and mapping techniques to gain a better understanding of public spaces in cities across the United States. [1]The project, led by Javier Argota Sánchez-Vaquerizo, was supported by the Metro 21 Institute and the Pittsburgh Downtown Partnership, studied Pittsburgh's emblematic Market Square, and resulted in a visual-data portrait of its urban activity. This portrait combined computational analyses and on-the-ground observations, and was both anonymous and deeply local. In contrast with the Lowry models of yore, this was not a predictive tool but an interpretive, and open-ended one. It helped trace fluctuations in the use of urban space in response to weather changes and to the disposition of urban furniture, suggesting new ways to study the relationship between built form and urban activity – as well as new questions about data access and literacy.

Rethinking urban form

Can critically-designed data structures and algorithms combine with other forms of interpretive research to enable new kinds of urban analysis that make visible long histories of urban change, and reveal spatial and infrastructural inequities, in a new light?

This project, developed by Jinmo Rhee at the Computational Design Laboratory, used deep learning, a subset of machine learning methods that leverages data representations and neural networks, archival research, and urban walks to shed new light on Pittsburgh's urban fabric. One of the project's outcomes is a novel kind of urban plan of the city that indicates variations in the use of public and private spaces, heights, and density: a high-resolution 'heat map' indicating the city's morphological gradients.[2] In this project, computational methods acted not to replace but rather enrich qualitative forms of observation and analysis. On the ground, through urban walks and through document analysis in the city's historical archives, Jinmo sought to corroborate and enrich the insights produced by his computational analysis. At a technical level this project facilitates a new kind of comparative analysis of urban fabrics within a city and across different cities. At a methodological level, it hints at an enriched toolkit for urban technology design that relies not only on the apparent trustworthiness of urban data but probes and situates these data critically alongside other forms of evidence, analysis, and experience.

Juxtaposing distant landscapes

Can platforms elicit new forms of co-presence that do not rely on production logics but enable unstructured interactions, new modes of creative engagement, and new understandings of the urban?

This project, an artistic collaboration between Andres Lombana Bermudez and myself, is a networked, audiovisual performance that links soundscapes and visuals from the two cities where we live – Pittsburgh and Bogota – captured during the COVID-19 quarantine period of 2020.[3] An exercise in juxtaposition, it places dissimilar sounds and imagery of ferrovial systems, water canals, urban fauna, and domestic life alongside synthetic soundscapes created using guitars, granular synthesis, and other software instruments. We performed part of the piece live using an online networked music performance (NMP) platform, and completed it asynchronously locally in our machines. The resulting piece, Paisajes distantes, hinted at the possibilities of new encounters, new forms of co-presence, and new kinds of creative engagement shaped by the necessity of isolation and by the affordances of online platforms and computational processes.