Dynamic Analysis Of Customer Transfer In Shanghai Business District Based On Big Data

The business district is the commercial sign of a city,and it also reflects the level of local economic development to a certain extent.Business district analysis can help national and local governments understand the development trend of business district,and provide a scientific basis for formulating business district development plans and policy orientations.In addition,business district analysis can also provide a basis for operators to choose business locations,formulate and adjust business policies and strategies.Based on the above background,this paper takes the Shanghai business district as the main research body,based on the data of UnionPay,and uses statistical methods to conduct an in-depth analysis of the problem which customer transfer consumption in the business district.Since the number of customer transfers in the business district is a time series of matrix data,and there is a high degree of consistency between the business district,the problem of singular matrix inversion will occur during the modeling process.Therefore,this paper first studies the multicollinearity problem of matrix data,and the criterion for multicollinearity is given.Then,it uses the multicollinearity discrimination method of matrix data,the business district with multicollinearity are eliminated,and the problem of singular matrix inversion is eliminated.At last,a matrix data time series autoregressive model is established for the business district data,and the law and reasons for customer transfer in the business district are analyzed based on the model.The two most important contributions of this paper are:First,Using the matrix data time series model to study the problem of customer transfer in the business district,the matrix structure of the data information can be well maintained,and compared with the study of straightening the matrix data into vector data,the model parameters are reduced,which is convenient for model analysis and discovery of the main laws of customer transfer problems in the business district.Second,a new multicollinearity test method is proposed.The advantage of this method is that the discriminant criterion has a clear geometric meaning,and the method can be used for multicollinearity test of matrix data.