Search

Abstract Time is a fundamental characteristic for understanding human activities. When analysing temporal pattern of a group of activities, most researchers tend to utilise one temporal attribute when representing time use of activities. Thus, temporal pattern of activities is usually visualised and understood as a profile of various observations listed sequentially over time. This paper aims to investigate the temporal pattern of activities in urban areas from a new perspective. Temporal pattern is visualised and analysed as the distribution of activity points in a two-dimensional temporal plane defined by the start and end time of activities as x and y axes. Kernel density estimation is used as a typical method to observe the temporal pattern of activities in Shanghai based on a one-week smart card dataset generated in the Shanghai's metro system. The results show that the proposed perspective can reveal considerably more information regarding the temporal pattern than a conventional one can. Highlights • We analyse the temporal pattern of activities from a new perspective. • Kernel density estimation is applied to observe the temporal pattern of activities using a big data approach. • Temporal patterns of activities are explored at different spatial scales. • Our new method is able to reveal more information about activities compared with the traditional one. [ABSTRACT FROM AUTHOR]