Research On The Location Of New Tea Stores Supported By Multi-source Spatial Data

Location-selecting directly affects the profitability and survival of the store.In the existing researches,store location is mainly divided into two kinds.One is GIS-based location selecting procedure,which is usually highly dependent on GIS software and uses a large number of spatial analysis tools.Although the effect is intuitive and easy to operate,it is still limited by the performance of the software itself and the limitations of the method itself.The other method is to use data analysis software like SPSS for modeling and prediction.This method leads to higher accuracy,but its algorithm is more complex in processing data.In recent years,with the development of the Internet,artificial intelligence,spatial big data and the popularity of LBS application,it is possible to use multi-source data in study of location-selecting problem.This article used the web crawler technology to get the source from public website open-source geographic data and business data,combined GIS software and machine-learning methods,to resolute the location-selecting of Shenzhen new-style tea store,mainly includes the following work: first,based on the results of the location theory and the actual situation of the research object itself,this paper analyzed the influence of various factors on site selection of the new tea shop,finally summarized as demographic factors,transport factors,competitive factors,other factors such as the four classes;Secondly use web crawler climb took gold POI data,Scott map data traffic situation,HOME LINK net second-hand housing prices data,data from public comments and the group's review,from the government's public platform for the urban renewal planning data,combined with the characteristics of the location factors used to train model is constructed,and use the SFS feature selection method for initial features set;Secondly,the location problem is transformed into a binary classification problem,and the classification algorithms commonly used in machine learning algorithms,including Logistic regression,support vector machine and BP neural network,are used to train the data set,and the model evaluation index system is combined to evaluate the performance of the model.Finally,based on the evaluation results of the model,the most suitable model for the location problem studied in this paper is selected,and the influencing factors of store location are discussed according to the results of feature selection and model training,and the improvement suggestions of the new tea store location model are put forward.