Font Size: a A A

Design Of Business Analysis System In Telecom Industry

Posted on:2010-06-26Degree:MasterType:Thesis
Country:ChinaCandidate:W Y ZhouFull Text:PDF
GTID:2178360278970769Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
With the reorganization of our telecommunications companies and the issuance of 3G licenses, the competition of telecom industry is becoming fiercer. Compared with other industries, telecom operators have more data of the customers. Those who can mine the knowledge contained in the data correctly can offer better product and service. It prompts telecom operators construct or improve the data warehouse-based business analysis system.On the depth analysis of the circumstances of data warehouse technology in the domestic and international telecommunications industry, this thesis points out the shortcomings of the existing systems, then points out the principles, objectives and steps of the construction of the system. On the basis of studying the related technologies of data warehouse and business analysis system, this thesis gives the detailed design of the structure and logic model of the business analysis system, Combined with the characteristics of the business processes of the telecom industry, this thesis divides the subject domain for the data warehouse, and gives the detailed model design of the customer subject which is the most important theme of this data warehouse. Combined with the characteristics of the data and the analysis requirements of this system, the granularity of the data is also designed too. This thesis designs the partition for the data according to the time and geographical position. This system applies a ECTL(Extracting\Cleaning\Translation\Loading) framework which is an optimized ETL framework. Data cleaning is joined in the ETL process to improve data quality. This thesis discusses data extraction, cleansing, transformation, loading and other related technologies. In addition, this thesis raises a method for the processing of the extreme data on the basis of distributive characters of the normal distribution and realizes this method by using storage processes. The application of this method has proved that this method could effectively mark out the extreme data. Then this thesis realizes the extraction process of Customers pay table by using PowerCenter.Finally, this thesis is summarized and points out what need to do next for this system in the future.
Keywords/Search Tags:Data Warehouse, ETL, Telecom Business Analysis System, Data Cleaning
PDF Full Text Request
Related items