Spatio-temporal Dynamic Analysis Of Urban Crowd Activity Based On Big Data

With the popularization of intelligent mobile data terminals and the development of internet technology, more and more social applications and location-based sevices begin to appear, following by this, a massive position data is also produced. Such location data includes special and temporal characteristics of these users’activities. Therefore, how to effectively use and analyze massive amounts of data, and extract and analyze spatial and temporal characteristics of the human beings from these data, it has become an important research topic. Now, the urbanization of our country has increasingly rapid pace, and its scale is further expanded, but also produces a series of "urban disease", such as the lack of urban land, the amount of urban traffic load surged, urban air pollution and so on. These are made suburbanization, and the phenomenon of jobs and residential isolated is became more common in big cities. The varying layout of employment and residencial areas causes the imbalances of working and living, and these problem are further intensified. By sensing and analyzing spatial and temporal dynamic changes of urban population activities, we can improve the most important data basis for taking advantage of resource coordinating the relation between men and nature for urban planning and management. The research idea of this article is through gathering user of Sina weibo location data in research area, analyze and mine the spatial location imformation of these data, from the different statistical analysis indicators we can obtain the spatial and temporal distribution characteristics of location weibo’s users in research area, and connecting with the land use types of the area, further analysis and mining the relationship between the group activities characteristics and the land use types. In this paper, the main research results are as follows:(1)This study chooses Sina weibo data and baidu map data as the research data, and discusses the current widely methods of the social network data capture. On this basis, through dividing the study area into 1 km* 1 km grids to improve the technical constraints problems in the process of fetching data. After checking the repeat and wrong data, the research obtained from the March to August Sina weibo user position data of more than 570000 article in 2014, four kinds of baidu map data POI points, a total of more than 2000. Because of the amounts of Sina weibo data, the research uses PostgreSQL database on storage and management of Sina weibo data. According to the late research train of thought, the research built a total of 47 database tables from the total to points about time scale. Baidu map data use Excel spreadsheets on storage and management.(2)According to the total to points of time scald, Sina weibo data is statistics into the grid, and through using the most value, the difference value, the spatial autocorrelation, the Mean Center, the directional distribution and the Kernel Density, six kinds of statistical analysis analyzing the differences of location weibo numbers in grids. The most value method get the maximum and minimum values of location weibo’s number in the research area girds under the different time granularity. The difference value method get the difference of location weibo" s number in the same grid under the different time granularity. Spatial autocorrelation analysis can obtain that the location weibo’s number in the study area has a strong spatial correlation, and analyzes the four kinds of local spatial autocorrelation agglomeration state. The calculation of Direction distribution (Standard Deviation Ellipses), which can come to from the perspective of macro research area position under different time granularity weibo data, namely the distribution of the position of weibo users and distribution and distribution direction. The calculation of average center (distribution orthocenter) can see the position of weibo users distribution orthocenter, the moving direction of distribution orthocenter in the different time. Kenel density calculation can get weibo users position distribution of intensity in the study area. Above the results of several statistical methods should combine with the land use types to describe the difference from the point of time and space, then it is concluded that the dynamic change characteristics of group activities under the different time scales in the study area.(3)According to the statistical analysis of the dynamic change characteristics of group activities, the relationship between men and nature in order to further study, by selecting the sina weibo data in June,2014, and combining with the location weibo data of POI type, compared the attractive intensity of the different types of POI points to human beings and the attractive intensity of the same types of POI points to human beings, thus further analysis the group activity information in grid.To sum up, under the background of the era of big data, this research is through the fetching the huge number of Sina weibo position data and Baidu map POI data, using Sina weibo API interface, Baidu map API interface, Locoy, ArcGIS, Geoda and other technical means, combined with many statistical analysis method, dynamic simulation and analyzing the spatial and temporal distribution characteristics of the location weibo users in study area, come to the relationship between the land use types and. It not only reveals the characteristics of dynamic changes of population in city, but also takes advantage of resource coordinates the relation between men and nature for urban planning and management.