Urban data and urban design: A data mining approach to architecture education

The configuration of urban projects using Information and Communication Technologies is an essential aspect in the education of future architects. Students must know the technologies that will facilitate their academic and professional development, as well as anticipating the needs of the citizens and the requirements of their designs. In this paper, a data mining approach was used to outline the strategic requirements for an urban design project in an architecture course using a Project-Based Learning strategy. Informal data related to an award-winning public space (Gillett Square in London, UK) was retrieved from two social networks (Flickr and Twitter), and from its official website. The analysis focused on semantic, temporal and spatial patterns, aspects generally overlooked in traditional approaches. Text-mining techniques were used to relate semantic and temporal data, focusing on seasonal and weekly (work-leisure) cycles, and the geographic patterns were extracted both from geotagged pictures and by geocoding user locations. The results showed that it is possible to obtain and extract valuable data and information in order to determine the different uses and architectural requirements of an urban space, but such data and information can be challenging to retrieve, structure, analyze and visualize. The main goal of the paper is to outline a strategy and present a visualization of the results, in a way designed to be attractive and informative for both students and professionals – even without a technical background – so the conducted analysis may be reproducible in other urban data contexts.     

Automatic tutoring system to support cross-disciplinary training in Big Data

During the last decade, Big Data has emerged as a powerful alternative to address latent challenges in scalable data management. The ever-growing amount and rapid evolution of tools, techniques, and technologies associated to Big Data require a broad skill set and deep knowledge of several domains—ranging from engineering to business, including computer science, networking, or analytics among others—which complicate the conception and deployment of academic programs and methodologies able to effectively train students in this discipline. The purpose of this paper is to propose a learning and teaching framework committed to train masters’ students in Big Data by conceiving an intelligent tutoring system aimed to (1) automatically tracking students’ progress, (2) effectively exploiting the diversity of their backgrounds, and (3) assisting the teaching staff on the course operation. Obtained results endorse the feasibility of this proposal and encourage practitioners to use this approach in other domains.