[2L5-GS-1] Fundamental AI, theory, algorithm:

[2L5-GS-1-01] Network analysis on human behavior using massive car GPS trajectory data

〇Haruki Sato¹, Yukihisa Fujita², Fujio Toriumi¹ (1. The University of Tokyo, 2. Toyota Motor Corporation)

Keywords:Data analysis, Global Positioning System (GPS), Spatial networks

In this study, we attempted to model the movement of cars as a network and to analyze its structure. The network is a large-scale bipartite graph network with vehicles and sites as nodes, based on the data of car trajectory data in Kaizu City, Gifu Prefecture, Japan. While large-scale trajectory data entails difficulties in visualization and quantitative evaluation, the network model enables analysis that can be visually and intuitively understood , and numerically evaluating structural characteristics. Networking also makes it possible to focus on the periodicity of people's movements and the role played by specific site nodes. The following two results were obtained as a result of this study. (1) Community detection within a bipartite graph allows for a good partitioning of living areas. (2) There are some sites that play a central role in the community as hub nodes of the network, such as supermarkets.