Research And Application Of Data Warehouse Based On SAP BW System

At present, China's power industry is vigorously promoting the construction of the information, and focus on development and utilization of information resources. Data warehousing is information management and analytical applications, an effective platform that can be more effective for the analysis of decision support system services, to improve the efficiency of their systems analysis and enhanced ability to handle complex queries. The dissertation focuses on the front line new research result on applied and the solution applied to Power Company. Most important, I apply SAP BW, which is the world leading BI system to the multi-dimensional analysis on management data, for extending the decision management information to high level managers through the SAP BW analysis tools. Main topics will include requirements analysis and theme design, data warehouse implementation, the key technology research.Based on status analysis of the business in electric power enterprise, their data warehouse platform is analyzed from nine major subject areas, which include financial management, marketing services, manufacturing operation, safety control, electric network management, planning and programming, project management, material management, human resources and organization management, through layer by layer decomposition of query from the various business units and KPI indicators, subdivide these into a number of themes on the basis of its nine business domains.This paper takes the theme of procurement analyzes as an example and detailed the process of data warehouse implementation, which includes establishing design ideas, clearing business requirements, modeling, data extraction, transformation and loading, data displaying. The project's design ideas clear for data source - operational data store - model three-tier architecture; business requirements are the analysis of purchase quantity, purchase amount, purchase price and the timeliness of receipt from multiple perspectives; business modeling includes the design of models' granularity, operational data store design and implementation, model design and implementation; the model data is from SAP R / 3 systems and business systems, which is cleaning, conversion and consolidation by the ETL tools of SAP BW system; in the model there are reports about purchase number analysis, purchase amount analysis reports purchase price analysis, receiving timely procurement analysis, which are showed via the WEB pages.The key technologies part discusses data loading technology solutions of SAP BW system, data integration solutions of electric power company, the research of data presentation solutions and data mining.Data loading follows the ETL process which extracts the source data from the source system, stores the data in data structures of the external system. The process is known as data extraction, through transport rules to achieve data conversion, through information packets and data transfer process to achieve the data uploaded. The overall architecture based on data warehouse system needs to extract data from about 10 kinds of business application systems, SAP BW system can not only easily from the SAP R/3 system to extract data, but also provides a database direct connection, SOAP, third-party data transmission, and other data extraction methods. Through analysis of a variety of data extraction methods, combined with the status of information systems in electric power enterprise, the paper presents the common file and database direct connection these two data extraction methods. Data presentation tools in SAP BW system contain a variety of reports, query design. It covers OLAP query, format of fixed form, dashboard analysis, which supports Excel, enterprise portals and third-party software (such as Crystal Report). OLAP processor in SAP BW system is rich in features, including navigation features (drill, rotate, slice), and filtering functions. OLAP analysis techniques are described through a example.Data mining research, firstly sets out the basics of data mining, including the implementation steps of data mining, CRISP-DM model and common data mining methods, and then makes a preliminary study on data mining in SAP BW system. In SAP BW data miningmodule, it can be categorized into two types of supervised and unsupervised learning.