Mining Place Design Knowledge from Multi-source Data in an Informed Design Platform

Along with the prevailing development of Smart Cities (SC), a novel knowledge management system is required to comprehensively manage data derived from ubiquitous physical and virtual objects for sophisticated and innovative analysis services. Particularly, in urban design, an "Informed Design" concept can be enabled by such system to harness big and heterogeneous geo-referenced data for a knowledge-based and responsive place design. As a solution to address emerging issues in collecting, storing, cleansing, analyzing, integrating, mining and visualizing multi-source data, an Informed Design Platform (IDP) is designed and being implemented. In this paper, an activity related multi-source data processing flow implemented in IDP is discussed to exemplify 1) how multi-source data can be efficiently and effectively processed based on concurrent data cleansing, analyzing and integrating cascades, 2) how indirectly related contents can be fully utilized to support a synergetic multisource data integration based on an outperforming ensemble-based social activity detection mechanism, and 3) how activity related place design knowledge can be easily mined based on a multi-dimension and multi-measure query (MMQ) to reveal place design context of the project testbed Jurong East, Singapore.