Community design of a light rail transit-oriented development using casewise visual evaluation (CAVE)

This paper proposes the casewise visual evaluation or CAVE, methodology and discusses its application to the participatory design of a transit-oriented development (TOD) in Louisville, Kentucky. CAVE is a fuzzy logic-based non-linear visual preference modeling system designed to provide design element guidance from composite visual scenarios under conditions of sparse data.The context of application in a low-income urban neighborhood is detailed. An architectural expert's design vocabulary allows model input and output to be structured. A small set of image samples was scored for preference using anonymous electronic polling in distributed neighborhood forums. Using fuzzy set theoretic software a community preference knowledge base (PKB) was built and interrogated. Four critical TOD design dimensions were selected: height, typology, density, and open space type. Preferred TOD design combinations were identified using the PKB and discussed. This project shows that CAVE can provide context-specific guidance for urban designers and that its strengths in effectively devolving design input and capturing local preferences are recognized by the community. The paper highlights the necessity for advanced geovisual analytic methods to be embedded into a structured public involvement (SPI) process.