As an external service window of power grid enterprise, the customer service center needs to integrate heterogeneous data sources within the enterprise, storage and analysis vast amount of customer service voice information and WEB service information. Hence, it becomes a very important topic on how to carry out the way to store various types of data in power grid efficiently, reliably, inexpensively and with availability of quick access and analysis. We have analyzed the key features and key technologies of big data, and designed the customer service data analysis system for power enterprises under big data Era. This paper covers data architecture, major system functions, key techniques and algorithms of the system. The system is designed on top of big data technology and data warehouse technology to centralized store, manage and use data, to achieve a unified data services platform through metadata management. The system is using Hadoop database for unstructured data storage, which works as data analysis and mining support platform as well. The data warehouse ETL component is designed based on CDC data warehouse ETL model, while the data presentation layer is using multidimensional data analysis techniques. Furthermore, the paper includes case review of the system which proves that the system is low cost with better scalability, reliability and parallel analysis features, etc. It may greatly improve the customer service level of power grid enterprise.