Bank Customer Credit Data Analysis System Based On Hadoop

The construction of credit is the cornerstone of the national economy.In recent years,China's big data,artificial intelligence and other technologies are showing a prosperous development trend,resulting in a number of innovative business models,such as mobile payment,P2 P online loans,Internet financial platforms and other new financial business models.However,there is no doubt that the Internet finance sector is still in its infancy.While the amount of data is growing rapidly,a series of problems have been exposed.Internet financial business types and workflows are becoming more and more complex,and the search for data values is becoming more and more in-depth.Traditional business intelligence analysis software is difficult to efficiently process massive and various forms of data.In order to solve these problems,a Internet banking customer credit data analysis system based on Hadoop is designed and implemented,which consists of four core parts,ETL(data extraction,transformation,loading),data modeling,workflow scheduling and data visualization.The research contribution has the following three points.Firstly,in order to improve the data processing efficiency of the credit data analysis system,a one-stop banking financial information analysis system based on Hadoop was designed and implemented.Compared with traditional data warehouses,HDFS under Hadoop architecture can support massive data storage.MapReduce can support distributed processing of massive data,and Hadoop-based data warehouse can support multiple data formats,such as pictures and videos.And the system can carry out efficient processing workflow through good scheduling design,which has strong practical significance in the actual production process.Secondly,in order to save space on HDFS,reduce data redundancy and perform data processing with higher efficiency.Applying the Data Vault model and the FS-LDM model in the construction data warehouse,by using the above model,effectively reducing the storage of duplicate access and duplicate data,can improve the operational efficiency of the data warehouse,and support business data expansion,saving data warehouse.Thirdly,the rating of the credit rating is achieved by using a model that builds a ten-level classification.From this,we can study and analyze the corresponding default rate of different personal characteristic data,establish a credit rating sys The construction of credit is the cornerstone of the national economy.In recent years,China's big data,artificial intelligence and other technologies are showing a prosperous development trend,resulting in a number of innovative business models,such as mobile payment,P2 P online loans,Internet financial platforms and other new financial business models.However,there is no doubt that the Internet finance sector is still in its infancy.While the amount of data is growing rapidly,a series of problems have been exposed.Internet financial business types and workflows are becoming more and more complex,and the search for data values is becoming more and more in-depth.Traditional business intelligence analysis software is difficult to efficiently process massive and various forms of data.In order to solve these problems,a Internet banking customer credit data analysis system based on Hadoop is designed and implemented,which consists of four core parts,ETL(data extraction,transformation,loading),data modeling,workflow scheduling and data visualization.The research contribution has the following three points.Firstly,in order to improve the data processing efficiency of the credit data analysis system,a one-stop banking financial information analysis system based on Hadoop was designed and implemented.Compared with traditional data warehouses,HDFS under Hadoop architecture can support massive data storage.MapReduce can support distributed processing of massive data,and Hadoop-based data warehouse can support multiple data formats,such as pictures and videos.And the system can carry out efficient processing workflow through good scheduling design,which has strong practical significance in the actual production process.Secondly,in order to save space on HDFS,reduce data redundancy and perform data processing with higher efficiency.Applying the Data Vault model and the FS-LDM model in the construction data warehouse,by using the above model,effectively reducing the storage of duplicate access and duplicate data,can improve the operational efficiency of the data warehouse,and support business data expansion,saving data warehouse's resources.Thirdly,the rating of the credit rating is achieved by using a model that builds a ten-level classification.From this,we can study and analyze the corresponding default rate of different personal characteristic data,establish a credit rating system through data processing,and grasp the degree of different personal characteristics corresponding to the default rate,which can be used for credit reporting.Report and guide the development of the risk control approval business.tem through data processing,and grasp the degree of different personal characteristics corresponding to the default rate,which can be used for credit reporting.Report and guide the development of the risk control approval business.