基于众源轨迹大数据的行为模式挖掘与定量空间优化

With the rapid economic development and urbanization, prominent contradictions stand out in the past decades, which is because of the space resources allocation shortage such as urban traffic congestion and the crowd stampede in Shanghai, and space resources idleness such as ghost town in Ordos. The key problem is that the disobedience of space resources allocation does not math the behavioral activity patterns of the crowd. Hence, mining space-time distribution patterns of the crowd from individual activity data, and optimizing the structure and function of space are the critical scientific issues to deal with the contradictions. .This project takes the media data in cyberspace (on-line) and activity traces in physical space (off-line) as data source. Identifying the spatial behavior as an entry point of this research, we build a space-time data fusion model of the on-line and off-line data, and put forward an innovative theory of individual activity identification and crowd behavior mining, based on space-time traces. The space-time distribution and evolutionary rule of the crowd behavior are unveiled as a result. We propose a quantitative space optimization approach based on space-time rule of crowd activity, which breaks through the bottleneck of detail activity data acquisition in traditional questionnaire survey. This project accomplishes an original innovation of space-time GIS in behavioral pattern analysis, space optimization, and overturns the dilemma of planning off the top of head, which contributes tremendously to the smart city building and fine city management.

随着经济快速发展与城市化进程，大量涌现空间资源配置不足（交通严重拥堵、上海人群踩踏事故）与空间闲置（鄂尔多斯“鬼城”）等矛盾，问题的关键是空间资源配置没有遵循人群活动规律，如何从个体活动数据中获取群体空间行为的分布，优化空间资源配置，是解决这些矛盾的关键科学问题。. 本项目以网络空间（线上）媒体数据与现实空间（线下）人群活动轨迹为数据源，构建线上线下空间轨迹数据的时空融合模型，突破传统问卷调查无法获取详细活动数据的瓶颈；以空间行为特征甄别为切入点，研究基于众源轨迹的个体活动识别与群体行为模式提取新理论，揭示人群行为活动的时空分布及演化规律；提出基于群体活动规律的定量空间优化方法，突破“拍脑袋”的传统“定性”空间规划，开启并发展基于人群空间利用规律的“定量”空间规划，实现时空GIS理论在行为模式分析、定量空间优化的原始创新，对智慧城市建设与精细化管理等具有重要研究和社会意义。

随着经济快速发展与城市化进程，大量涌现空间资源配置不足（交通严重拥堵、上海人群踩踏事故）与空间闲置（鄂尔多斯“鬼城”）等矛盾，问题的关键是空间资源配置没有遵循人群活动规律，如何从个体活动数据中获取群体空间行为的分布，优化空间资源配置，是解决这些矛盾的关键科学问题。. 本项目以网络空间（线上）媒体数据与现实空间（线下）人群活动轨迹为数据源，构建线上线下空间轨迹数据的时空融合模型，突破传统问卷调查无法获取详细活动数据的瓶颈；以空间行为特征甄别为切入点，研究基于众源轨迹的个体活动识别与群体行为模式提取新理论，揭示人群行为活动的时空分布及演化规律；提出基于群体活动规律的定量空间优化方法，突破“拍脑袋”的传统“定性”空间规划，开启并发展基于人群空间利用规律的“定量”空间规划，实现时空GIS理论在行为模式分析、定量空间优化的原始创新，对智慧城市建设与精细化管理等具有重要研究和社会意义。