大数据行为研究趋势:一个“时空关”的视角

本文从"时空关",即时间、空间、关联3个维度来分析近十余年大数据行为研究的进展,总结近年来本课题组的相关研究结果,并探讨未来发展趋势。各种时空大数据的广泛采集和应用取得了突破性的进展,推动了有关人类流动行为的研究;与此同时,社会网络和社会媒体大数据的发展把人们关联行为的研究推向新高度。流动性和关联性以及这两个方向的交融构成了大数据行为研究的基本范畴,而通过"时空关+"不同内容和场景形成了大数据应用的广泛格局。大数据行为研究面临的主要挑战是如何解决大数据分析中的内生性问题,未来的研究将更加重视通过随机实验、田野实验或者自然实验等方法来做因果推断。大数据行为研究的"时空关+"为探讨人类行为的基本规律以及各种应用场景提供了一个崭新的视角。

Big Data Behavioral Research Trends:A Time-Space-Connection Perspective

We use the three dimensions of time,space,and connections(TSC)to analyze and interpret big data behavioral research progress in the last decade,to summarize our own related research efforts,and explore the future research trend.The widespread collection and application of spatiotemporal big data in a wide set of technological,commercial,and social contexts has seen breakthrough progress and promoted human mobility behavioral research.At the same time,the big data based developments of social networks and social media has pushed the research on human connections to new heights of importance and sophistication.Mobility and connections as well as their interactions form the basic behavioral support and foundations for big data research;TSC+content and contextual extensions also provide broad pathways forward for big data applications.The major challenge for big data behavioral research is how to solve endogeneity problems in big data analysis,and future research are likely to focus on experiment-based causal inference,using randomized field and natural experiments.The TSC+perspective of big data behavioral research provides a new framework for studying human behavior regularities and a wide variety of applications.