Understanding temporal pattern of human activities using Temporal Areas of Interest

Although researchers have realised the importance of time for decades in geography and relevant disciplines, they more or less ignore how our understanding can be greatly enriched through the lens of time. One outcome of this ignorance is that there are a limited number of concepts and methods to analyse the temporal pattern of human activities. In this paper, we try to understand the temporal pattern of activities by simultaneously examining the start and end time of activities, rather than from a conventional view, which usually understands a temporal pattern of activities as a profile of an attribute (e.g. the number of activities) on the time axis. To achieve this, we propose a new concept, Temporal Areas of Interest (TAI), defined as the area on a two-dimensional plane using the start and end time as the X and Y axes, where activity points are densely distributed and therefore attracts people's attention/interest. We also propose a general methodological framework for identifying TAIs. The concept and framework are applied using a week's metro smart card data in Shanghai. Results show both the basic rhythm of human activities and reveal some facts that are unexpected or ignored before, such as the large existence of overtime work or similar activities on weekday evenings and weekends. We think that the proposed method provides an alternative way to understand the temporal pattern of activities and a relatively new perspective in evaluating human activity patterns and the urban functions across the city.