One key role of urban informatics is to understand demands, challenges, and opportunities for major infrastructure projects in a bottom-up format. In this respect, engaging citizens through communication tools available in Social Web, and distilling micro-knowledge distributed among them appear to be reliable and promising streams. Chaotic nature and large size of unstructured big-data gathered in this way, and semantic barriers for understanding and classifying the content are among other difficulties which challenge efficiency and effectiveness of Such methods. This paper suggests a specific form of collective intelligence, which can be described as 'collective classification', as a strategy and employs crowdsourcing as a tool to overcome such difficulties. We explain crowdsourcing in form of a Game With A Purpose, designed and run by University of Toronto between June and August of 2013. The game aimed to classify the content of online feeds in the context of sustainability of infrastructure projects, and to form a folksonomy for decoding the future communication. Players of the game were provided with 'sustainability' (at a holistic view covering its three main pillars: economy, environment, and society) as the context, a set of relevant tweets as items to be annotated, and a set of indicators developed in the domain literature as a template to follow. On top of the predefined indicators, the participants were allowed to suggest and create their own classes. Players' inputs were used in scoring other players' performance and this refined the results by drawing on the 'Wisdom of the Crowd'. The core contribution of the research is to use human computation in a bottom-up approach to solve a transdisciplinary problem which traditionally has been tackled via top-down solutions. As an example, taxonomies created by domain professionals are normally biased towards professionals; such traditional solutions normally lack context-sensitivity in nature, constructivism in approach, and pluralism in perspective. The method and the platform presented by this paper can be directly applied in similar studies to benchmark, classify, and fine-tune public perspective/knowledge in urban infrastructure related technical or semi-technical issues. (C) 2015 Published by Elsevier Ltd.
