一个基于基站轨迹数据的城市移动模式可视分析系统

Trajectory2Vec: A Visual Analytics Approach for Urban Mobility Patterns Based on Mobile Phone Data

摘要: 随着移动通信技术的发展,手机基站轨迹数据在分析人类移动规律方面的优势日趋显著.由于人群移动模式与其社会行为息息相关,该模式能够直接反映各地理区块在不同时间段所具备的社会功能.根据词嵌入模型,首先将基站的时空信息映射为向量,通过计算基站间的高层语义的相似规律来分析地理区域的功能性信息;再将带有时空变化信息的手机用户移动轨迹映射至向量空间,使基站地理坐标与轨迹相结合,从而获取更加丰富的语义信息.在交互方面,设计了一个可视化分析系统Trajectory2Vec来探索城市区域功能和用户行为的关系,案例分析证明了该系统可以有效地帮助用户分析移动人群与城市区域间关系的动态变化规律.

Abstract: Based on word embedding model,we map the spatio-temporal information of base stations to the vector space and calculate the similarity rule of the high-level semantics between base stations to analyze the social information of geographical areas.Moreover,we designed and implemented a visual analysis system to explore the relationship between urban interregional functions.Case studies show that the proposed system can effectively help users to analyze the dynamic changes of the relationship between urban area and local residents.