Search

Smart card data have received increasing interest from transport researchers for travel behavior investigation. However, the utility of smart card data has arguably yet to be exploited in the spatio-temporal context of interpreting the essential properties of passenger groups. Before building socially responsible data models, considering ethics issues of what is right, just and fair is able to minimize negative impacts of technology. This work demonstrates a multi-phase methodology to illustrate the essential properties of passenger groups with crowd mobility portraits that are presented by several words in combination with visualized maps. In particular, we propose an efficient ranking algorithm to figure out trajectories mostly contributing to the city-wide mobility patterns and construct time-varying land functions to textualize trajectories and obtain interpretable crowd mobility portraits. Experiments show that our method outperforms the state-of-the-art methods with an accuracy improvement of 8%. Case studies of real-world data also confirm the effectiveness of crowd mobility portraits. [ABSTRACT FROM AUTHOR]