Abstract: 【Purpose/significance】Social sensing is the theory and method of studying the characteristics of human spatiotemporal be?
havior with the help of massive spatiotemporal data, and revealing the spatiotemporal distribution, connection and process of social and economic phenomena. User profile aim to reveal social phenomena by mining user attribute characteristics and behavior patterns,
which is an important means to achieve social sensing.【Method/process】Around the three dimensions of emotional cognition, behav? ioral habits and social network relationships, sort out the content dimensions of user profile.【Result/conclusion】The study found that multi-source heterogeneous spatiotemporal data can be used; the research content focuses on emotion and place semantics, spatial in? teraction, and social network mining; spatiotemporal semantic reasoning is a new research area; the application scenarios can be ex? panded in cross-regional spatial information services, mining of urban human-land interaction, efficient government governance, and prediction of spatial-temporal correlation trends of domain events.【Innovation/limitation】By integrating multi-source heterogeneous spatiotemporal data, accurate knowledge services in multiple scenarios can be realized.