Partially supported by a NULab Seedling Grant.
The next 40 years will see more urban fabric built than we have in the whole of human history. This growth exceeds existing housing provision and urban infrastructure, tending instead towards informal, unplanned communities. Occurring without the work of architects or planners, this informal growth is also further perpetuating existing structural inequities in cities and social systems. The mismatch between the scale of contemporary infrastructure and that of the typical city-dweller represents a gap in participation, ownership, and engagement.
Informal settlements are characterized by informal land tenure and unplanned land uses, which oftentimes results in lack of physical access to infrastructure and services, including water and sanitation. Re-blocking is the process of formalizing land tenure and providing street access to each parcel and building in a neighborhood to facilitate the introduction of piped urban services and minimize the implementation and maintenance cost for local governments providing the services. However, re-blocking can be highly contentious due to the disruptive implications of the interventions, the absence of tools to map these neighborhoods, the difficulty of coordinating numerous stakeholders with the different cost/benefit tradeoffs of their designs, and the inability to make sense of and synthesize multiple plausible resident proposals.
Responding to such challenges, this research project aims to collectively map and characterize thousands of informal communities across the world and to explore the use of collaborative design tools in community design and deliberation practices to create cadastral maps for these communities. Such collaborative design tools will help communities synthesize, evaluate, and modify synthetically generated re-blocking proposals in a scalable, contextual, and place-based manner.
This work leverages computer vision and generative AI to create baseline cadastral maps for data scarce urban environments at scale. This project will continue the development of an image-to-image generative AI model to create plausible parcels maps based on building footprint vector data that our team has created based on satellite imagery.
The second stage of the research will leverage the thousands of parcel maps generated by our model, to serve as a baseline proposal that residents can modify and re-imagine though a collaborative online platform. This platform will enable residents of informal communities to collectively modify the maps based on their contextual knowledge and interests.
This work investigates how the synthesis of community collaboration, data, and computing can support community-developed interventions and the types of deliberative design practices that result from such processes. This re-blocking project will provide a proof of concept to envision the ways in which new models of collaborative computing and community organizing can promote a more emergent and community driven approach to architecture and urban planning that can support future challenges in urbanization; it inquires into whether computing can play an ethical role in such transformation.
Principal Investigator
Carlos Sandoval Olascoaga, Assistant Professor, School of Architecture, CAMD
