Telcoflow: Visual Exploration of Collective Behaviors Based on Telco Data

Yixian Zheng,Wenchao Wu,Haipeng Zeng,Nan Cao,Huamin Qu,Mingxuan Yuan,Jia Zeng,Lionel M. Ni

DOI: https://doi.org/10.1109/bigdata.2016.7840677

2016-01-01

Abstract:Collective behavior is an important concept defined to capture behavioral patterns emerged among the crowd spontaneously. In social science, people's behaviors can be regarded as temporal transitions between a set of typical states (e.g., home and work) which are always associated with certain locations. This fact leads to an interesting research topic in developing ways to explore people's collective behavior patterns through movement analysis, which is our focus in this paper. In recent years, massive volumes of spatio-temporal data generated by mobile phones, called telco data, bring an unprecedented opportunity to study collective behaviors in terms of large coverage and fine-grained resolution. However, distilling valuable collective behavior patterns from the large scale of telco data is not an easy task. The challenge is rooted in two aspects, including the data uncertainty as well as the lack of methods to characterize, compare and understand dynamic crowd behaviors, which triggers the use of visual analytics to take full advantage of machines' computational power as well as human's domain knowledge and cognitive abilities. In this paper, we propose TelcoFlow, a comprehensive visual analytics system which incorporates advanced quantitative analyses (e.g., state-based behavior model) and intuitive visualizations (e.g., an extended flow view embedded with state glyphs) to support an efficient and in-depth analysis of collective behaviors based on telco data. Case studies with a real-world dataset and expert interviews are carried out to demonstrate the effectiveness of our system for analysts to gain insights into collective behaviors and facilitate various analytical tasks.