Identification and Prediction of Urban Travel Hot Spots with a Potential-Based Approach: A Case Study for New York City

Identifying and predicting urban travel hot spots are crucially important for urban planning and safety management. Recently, the increasing availability of taxi data contributes to the exploration of travel characteristics of inhabitants. In this study, the taxi data of New York City is employed to generate the vector field of human mobility. The key property of vector field potentials is used to quantify the attraction of different locations and identify urban travel hot spots. In addition, the normalized potentials were used to generate the Markov chain model and predict urban travel hot spots. Experimental results indicate that it is feasible to identify urban travel hot spots with the potential-based method, and urban travel hot spots can be predicted using the Markov chain model with normalized potentials as inputs.