Consilium Urbis---A collective intelligence approach for perception-based urban analysis.

摘要

The research presented within this thesis is intended to provide a new analytic method for examining the city. It suggests that some of the existing challenges of traditional methods can be overcome with a computational methodology for perception-based urban analysis. Our project is motivated by the observation that a globally increasing rate of urbanization provides new challenges for urban planning, urban design as well as for policy decisions, while innovation in information technology offers potential for new conceptualization and advancement of traditional analytic methodology.;Our research shows how former approaches for spatial representation and urban analysis can effectively transition to digitally enabled feedback systems for reporting on urban conditions. The contemporary progression of geospatial technologies facilitates new approaches for assessing and evaluating urban processes which were not accessible in the past.;In order to develop an analytic framework for understanding the urban environment from the perspective of human perception, a novel method for cognitive association is introduced. This model provides a conceptual differentiation between perception-based and cognition-based criteria for urban analysis. Our research proves how such methodology generates a high degree of analytic granularity as well as the opportunity for translation into algorithmic models. It also shows that methods of computational collective intelligence allow for fusion of human and computational intelligence by generation of a shared information commons on the one side and by application of information foraging and filtering techniques to extract relevant analytic information on the other.;In consequence, an algorithmic model is established and implemented within a research prototype for experimental testing of various analytic aspects. The system resembles a publicly accessible collective intelligence platform which provides a collective information commons. Our experiment tests our approach for parameters such as individuals' cognitive and perceptive similarity, knowledge, cognitive propensity and cognitive confines, as well as demographic characteristics.;By applying empirical evaluation methods from machine learning theory and quantitative analysis, we are able to test the validity of our conceptual and technological approach for perception-based urban analysis. It is our hope that our findings might lead to an improved knowledge how to plan, design and experience future cities.