Visual Analysis Of Restaurant Consumption Behaviors In Urban Based On Data Of DianPing

With the development of internet and information technology,shopping mode like group buying online rise rapidly and generate huge amounts of multidimensional and heterogeneous data which reflect the features of urban restaurant consumption behaviors.This article discusses the visual analysis of urban restaurant consumption behavior based on online group purchase data in three aspects: Urban restaurant consumption data mining based on data of DianPing;Visual analysis system frame of urban restaurant consumption behavior based on multi-view and multi-granularity;Visual analysis system for urban restaurant consumption behavior based on Django.(1)Urban restaurant consumption data mining based on data of DianPing.According to the basic attributes of the data and high-level semantic features,a catering consumer behavior model was constructed from two perspectives:community-oriented and individual-oriented.In consumption satisfaction mining based on sentiment analysis,firstly,the supervised machine learning algorithm is used to categorize the sentiment of the review text.Then the consumption satisfaction calculation method is used to extract consumption satisfaction.In consumer groups with same consuming habit mining based on community detection,a method of building consumer network based on online group purchase data is proposed.Experiments prove that this method can effectively identify the consumer groups with similar consumption characteristic.(2)In visual analysis scheme design of urban restaurant consumption behavior,a sentimental analysis bubble chart was designed by spatial expansion and color mappings,as the sentiment are affected by many factors.A visual analysis view based on enhancement of location information is designed,as the sparsity of individual consumption data in spatial.The visual layout scheme based on collaborative and interaction of multi-view is proposed according to the diversity of analysis tasks.A datafiltering method based on multi-granularity in time-space dimension are proposed as the complex and large dimension information of restaurant consumption data.(3)Visual analysis system of urban restaurant consumption behavior based on Django.Visual analysis subsystem for urban restaurant consumption behavior for group and individual were developed as the analysis of variety of objects.The regional characters,consumption satisfaction characteristics,topic of consumption,key words,the Spatial and temporal correlation characteristics of regional sales of group restaurant consumption behavior were extracted,the taste and location preferences,time sequence and consumer group with similar consumption habit of individual restaurant consumption behavior were extracted too by the system and the 30572 reviews of DianPing in 11 months.The effectiveness of this system was verified.