Design And Implementation Of ETL Based On The Telecom Customer Care Project

To effeciently manage the telecom stock customers, Telecom Customer Care Project (CCP) has been raised as the analysis system which is built on the base of subordinative data mart, ETL is the process of data extracting, transformation, cleaning and loading, and is the most important step in data mart, and is the key of CCP analysis.This paper's main content is the design implementation of ETL process based on CCP, and the optimization of the ETL. The main work is as follows:First, study the concept and character of data warehouse, data mart and ETL for the theoretical fundation. Conduct detailed data analysis to get ETL processing structure on the basis of analyzing customer care project needs.Then, using E-LT system structural design in order to improve performance , ETL was designed to deal with the functions of data extraction, conversion cleaning, loading and data monitoring module, joined the log processing functions, in order to facilitate monitoring the entire treatment process, thus get the ETL processing framework. For the treatment for conversion of data quality problems, the paper raised the data quality assured ETL processing method based on data understanding, tracing and analysing the processing data and transformed data of all steps, preparing the corresponding cleaning conversion rules, and thus fully guarantee the quality of data. And ETL processing was divided into six serial processing of the independent testing of ETL processing module so as to facilitate unit testing and maintenance of expansion.Finally, the paper implemented ETL scheduling process from the overall to partial, achieved the processing modules function and ETL processing log function, and to the problems encountered by ETL realized process, used a gradual isolation of the test method to performance optimization. The functions and performance have reached the design requirements, provided appropriate data supports for CCP.