A Laplacian Evolution Approach to Identify the Source-Sink Pattern of Human Mobility in Urban

The comprehension of human mobility patterns can offer valuable insights into the factors influencing human movements and the relationships between individuals and the structural layout of urban areas. This can have positive impacts on diverse areas including urban transportation, urban planning, urban policy development, and public health. In this paper, we introduce a straightforward Laplacian evolution method to detect the source-sink pattern of human mobility within urban areas. Specifically, we calculate the divergence of the human flow field, which is equivalent to applying Laplacian operator to the human mobility density, to identify sources and sinks in urban areas and analyze human mobility patterns from it. This paper selects four typical areas as experimental subjects, including business district, school, industrial area and residential area. And a week’s worth of mobile signaling data obtained from Quanzhou, Fujian Province, China, was used to evaluate the human mobility patterns of the experimental subjects. The results of our analysis indicate that human mobility patterns, irrespective of the type of area, can typically be classified into two modes: one for working days and another for rest days. The characteristics of each mode are consistent with conventional knowledge, suggesting that our research participants have, to a certain degree, fulfilled the urban functionality expectations. This further confirms the efficacy of the methodology we have put forward. Our implementation is available at https://github.com/zpnightmare/flow.