Research On Urban Landmark Extracting And Commercial Area Mining With Check-in Data

As the most centralized and the most active area of material civilization and spiritual civilization in human society, urban is the necessary space for urban residents’ daily work and life. Urban spatial layout refers to the geo-spatial location, spatial distribution characteristics and organization structure of each functional area in city, is the urban functional arrangements’projection in space. The urban spatial layout information, which is an important part of urban spatial layout, mainly refers to the spatial distribution characteristics of each functional area in city. As the important reflection characteristics of urban spatial layout information, the urban landmarks extracting and commercial areas measuring can significantly improve the layout of city planning and promote the city economic development. However, the existing methods of urban landmark extracting and commercial area measuring often need to collect a large amount of data, which would cost a lot of money and energy, so that they can be adapted for only a small scale. Meanwhile, determining the values of some parameters in these measuring methods may be effected by human factors, which may induce subjective and uncertainty measurement results.The emergence and booming of crowdsourcing geospatial data provides a possible way to extract urban spatial layout information fully and accurately, such as urban landmark, urban commercial area, etc. As compared with conventional geographic data collection and update method, the crowd sourcing geographic data from the non-professional has characteristics or advantages of large data volume (data can be updated and edited by multi-people in many times), high currency (data can be updated by user the first time), abundance information (there are lots of social attributes in the data) and low cost. In this context, regarding crowdsourcing check-in data as the researching objects, by spatial statistical analysis and spatial data mining methods, the urban landmarks extracting and commercial area measuring methods based on crowdsourcing check-in data were introduced in this study. Focused on the approaches of landmark extracting, commercial area mining and commercial area analysis, the main research contents and innovative results are as follows:A quality analysis model for crowdsourcing check-in data is proposed. Firstly, a quality analysis framework is designed based on data characteristic analysis of check-in data. Secondly, a quality assessment model for check-in data by three different quality elements:data completeness, attribute accuracy and positional accuracy is presented. Finally, take the check-in data of Wuchang district in Wuhan for instance, the quality of check-in data is analyzed and assessed with 2011 version of navigation map for reference. The result shows that the check-in data features rich contents and high currency, but its weakness lies in poor standardization and positional accuracy. Moreover, the check-in data reflects regional distributions of urban economics in some extent so that it can be used for data mining research in urban landmark extracting and commercial area measuring.A method used to extract urban hierarchical landmarks from check-in data is proposed. Firstly, a POI significance measure model is constructed after analyzing the factors influencing the significances of POI objects from three vectors which are check-in numbers, check-in users and user impact factor in check-in data. Secondly, the Voronoi diagrams based on POI significances are constructed to segment the POI space scopes. By neighborhood analysis and merge algorithm of Voronoi diagrams iteratively, the urban landmarks are extracted hierarchically. Finally, an experiment is carried out to extract urban hierarchical landmark based on POI significance and Voronoi diagrams with check-in data in Wuchang District and Beijing. The experiment confirms the efficiency of this algorithm, makes further analysis on the regularity and spatial phenomenon reflected by timed hierarchical landmarks.A method used to extract urban hotspot and commercial area with check-in data is presented. Firstly, a check-in data grid-process model is proposed to simplify the number of check-in data and improve efficiency of cluster analysis. Secondly, the spatial autocorrelation is carried out to validate the spatial pattern of check-in data and find the most optimal distance for this pattern. Thirdly, an exploratory spatial analysis and hotspot clustering method based on check-in data is proposed to detect urban hotspot and extract commercial area by geographic distribution metric with urban commercial hotspots. Finally, an experiment of urban hotspot detecting and commercial area mining with check-in data in Wuhan is carried out, and then the mined commercial area information are analysed.A method used to analyze the urban commercial area morphology dynamically with check-in data is presented. Taking urban commercial area as a main object of study, a commercial area morphology dynamic analysis model is constructed from three vectors which are time morphology of commercial area, spatial morphology of commercial area and commercial area scale. Then, the dynamic evolution and developing trends of commercial area morphology are analyzed after extracting commercial areas respectively with check-in data as of 2011-09 and 2012-09 in Wuhan. The analysis result shows that the economic development of Wuhan have been maintain good and urban commercial areas keep further growth and expansion, while some measures like commercial diversification are necessary to be implement to prevent the recession of individual commercial area.