| Data has been generated at higher and higher speed. Data centers were developed to provide decision making support, such as reporting, subject ananlysis, data ming, etc. However both the technology and business pattern has been changing way too fast, resulting in the frequent modification in either business process or data entity which bring challenge to data warehouse to update the data model and platform in time.To not suffer the change, data platform must be designed with higher ease-to-use, higher flexibility, and higher extendability.A cross country eCommerce company has built an data platform infrastructure to performe configurable data extraction, transformation and loading. Main disadvantages of this infrastructure are:1, Data are simply replicated and be lack of integration.2, Data analysis is performed on raw table or department-dedicated data marts, resulting problem such as indigestion, bad query performance, long development cycle, poor data management. The ECAP, eCommerce Analytic Platform, was developed to solve the problems above.This thesis designs and implements the data model of ECAP. Dimensional modeling can provide the same information as3NF model via a format that-provding ease to understand, high performance to query and flexibility to modify. Different kind of analytical requirements can be met via different data design in each tier hence the data warehouse can be well managed. This thesis firstly analyzes the requirement of ECAP, then discuss the selection of business process, granularity design, dimension design. At last, the thesis demonstrates the uasage of dimensional model via SQL query and Bl tools to show the intelligibility, better query performance and extendability.The ECAP adopted dimaesiional data modeling and tiered data architecuture techonology solves the problem above.Then so-called Agile Analysis can be run in modern eCommerce company since user can concentrate on the business model rather than the detail of technology... |
PDF Full Text Request