Crowd-sourced urban life monitoring: urban area characterization based crowd behavioral patterns from Twitter

Location-based social network sites are recently attracting a great deal of attention by combing Web-based social network and the real-world location tagging in an integrated way, where people can publish their life logs about their real-world activities and share them with the public often looking for location-based information. Obviously, in terms of technological and social advance such as location sensing smartphones, experiences and thoughts by the unexpectedly growing number of the mobile users in urban area are conveniently being shared significantly impacting our ways of life experience sharing. In such context, we are able to monitor crowd's experiences through the location-based social network by collecting and analyzing crowd's numerous micro life logs to support a variety of decision makings. In this paper, we attempt to look into the crowd's urban lifestyles, which are characterizing urban areas, particularly utilizing Twitter. We provide a model to construct systems for a large-scale urban analytics with the location-based social network. We also describe our practical approach to describe urban characteristics represented by crowd's temporal behavioral patterns. In the experiment, we show an urban characterization by way of crowd's behavioral patterns, which are derived from temporal patterns of crowd behavior indirectly speculated from a massive number of collected Twitter messages. Finally, we discuss the importance of this kind of challenge amid the pervasive social network environment and some critical issues to be considered for the wide spectrum of sociological studies requiring technology-driven crowd life monitoring.