Research And Application Of LBS Spatial Data Clustering Based On MapReduce

LBS is based on location service is the English abbreviation, in recent years because of the rapid development of the positioning technology and Internet, the application field is more and more big, cost is getting lower and lower. LBS data, especially LBS data about the vehicle, is an objective description of the traffic of the city. Through the LBS data clustering analysis can be qualitative and quantitative analysis of the different areas of the city in different time traffic flow, vehicle speed, congestion information, and from different angles to describe the whole urban traffic characteristics.In this paper, the parallel computing MapReduce framework as the basis, of massive LBS trajectory data clustering analysis, in order to identify the relevant characteristics of the urban traffic, and put forward a method for effective analysis of large scale spatial data.The main work of this paper is as follows:1. Discuss the LBS data composition, structural features as well as the domestic and foreign current application situation of LBS is described, and the use of Apache Hadoop distributed parallel computing the MapReduce framework and realizes the space of LBS track log data aggregation.2. Compares and discusses the similarities and differences between clustering and spatial clustering, and for spatial clustering characteristics, and puts forward some be able to juggle a spatial relation of the spatial relationship and attribute information of the concept and the application of the model.3. Implementation of the clustering of more than one terabyte mass point data, in the paper presents a use cluster distributed computing to achieve data aggregation, and extraction method of statistical feature of the corresponding region of space, a large number of discrete point data aggregated into with the corresponding characteristics of unstructured grid data. Finally, the clustering is realized by the method of face shape spatial clustering in order to identify the characteristics of urban traffic.4. The design and coding to achieve a large number of LBS data analysis of the traffic situation of the application software, and in the actual work has been tested.