The spatial interaction network is the space-embedded directed network with flows of people, goods or information among places. Examples of spatial interaction networks include the international trade network, the migration network, the transportation network and the inter-city telecommunication network. Understanding and predicting spatial interaction patterns of these networks are of importance in various disciplines, including the regional science, the transportation science and the economic geography, with many practical applications in the urban or transportation system planning, the epidemiology of infectious diseases, and the location-based services. This paper first introduces briefly two classic spatial interaction models:the gravity model and the intervening opportunity model, and then reviews some recent advances in data-driven spatial interaction models, including the radiation model, the population-weighted opportunity model and the random walk models on spatial interaction networks. This paper also discusses some challenging problems in modeling the spatial interaction networks, such as the individual mobility behavior diversity, the group choice decision behavior experiment and the datadriven modeling approach for the human activity-travel behavior.

spatial interaction  /  complex networks  /  human mobility patterns  /  trip distribution