Spatial-temporal Distribution Analysis Of Emotion Based On Microblog Geography Data

Emotion is a kind of geography knowledge existing in space and time,but its acquisition is difficult.Emerging location social networks provide a good source of data for emotional metrics.At present,the sentiment analysis technology for short texts in social networks is still at a stage of development,and a unified and mature theoretical system has not been formed.With the "turn of emotions" in Western geography,the analysis of the emotional characteristics of Chinese microblogs has gradually become a new focus.This article is based on approximately 3.45 million microblogs with geographies released on September 19-25,and October 1-7,two weeks,using a combination of sentiment analysis and GIS methods to analyze the temporal and spatial characteristics of group emotions.Through text data preprocessing,Chinese word segmentation,and sentiment analysis methods based on dictionaries and grammar rules,the sentiment tendencies of each microblog message are scored,combined with the geographic location and time of the microblog,and the microblog emotional spatiotemporal feature vectors are constructed and visualized.The research results based on microblog data in two periods of weekdays and holidays throughout the country show that the number of microblogs released in China is more in the east and less in the west,and the overall emotional tendency is mainly active,but the spatial distribution on the individual level is random.This paper aggregated emotional attributes in cities,then analyzed the spatial distribution characteristics of urban population emotions,and detected the spatial correlation of urban emotions based on Tobler's first law of geography.The group sentiment scores based on cities show a pattern of subregional distribution in space,and the emotional attributes of neighboring cities in space have a similar tendency.Through comparative analysis of the local spatial self of emotional scores in two time periods.Relevance,we found that the local differences in the spatial distribution of sentiment scores among urban residents in China,the workday more significant than the holidays,detected more clustering and abnormal aggregation areas.In order to detect the distribution pattern of emotion in time and space environment,the time-snapshots are used to analyze the urban emotional space and temporal changes.The time-space cubes are used to aggregate the Weibo emotion data into point data sets.The time-space hotspots and anomaly detection methods are used to identify the hot spots,cold spots and outliers.The results showed that at each location,the sentiment scores in September were mainly increased while the national sentiments declined more during the National Day.The results show that there is a large gap in happiness between the eastern and western regions of China.In the eastern region,especially Shandong and Jiangsu,the pattern of emotional hotspots in the space-time neighborhood appears,while in the western region,especially in Xinjiang and Tibet,it is the emotional cold spot model.Prominently,the types of outliers in most areas will change over time,and at the same time,the key time steps of outliers has been detected.The interaction of the virtual social network is mapped into the real geographical space,an emotional interaction network based on the user's location interaction is constructed,and an emotional interaction index is introduced to analyze the intensity of the emotional interaction between cities.It is found that the interaction of information flow in social networks exists in the "rich club" phenomenon.The main emotional interaction chain is distributed among the cities in the central and eastern regions of China,especially in first-line developed cities such as Beijing,Shanghai,Shenzhen,and Guangzhou.The network occupies an emotionally oriented position.The conclusion of this paper reveals the diversity and difference of the spatial and temporal distribution of group emotions.Compared with traditional questionnaires,statistical data,and other methods,Weibo data has the advantages of large sample size,strong timeliness,and low cost,providing a method for emotional geography research.And it has guiding significance for analyzing the well-being of Chinese residents and supporting social and economic development planning.