Rethinking the Urban Design Process from a Data Perspective

Abstract

Urban design always requires the processing of large amounts of data from multi-disciplinary sources during the decision-making stages. However, unfamiliar multi-disciplinary data sets can only lead to confusion and uncertainty. This research proposes a data-driven approach for supporting the urban design process. A hybrid data mining method is used to cluster, classify and rank solution-instances according to geometrical properties and energy performance. An urban design case study is used to demonstrate the proposed method with respect to two performance issues: solar heat gains and natural ventilation. The result shows that the method addressing both familiar and unfamiliar data can effectively guide the designer during the design process.