Bi Modeling Principles and FAQs

The quality of BI modeling directly affects the quality of the Data Warehouse project, so there is a clear definition of the structure, size, and model capabilities of the data Warehouse before modeling.

There are many factors that affect the modeling of BI Data Warehouse, which often change depending on the specific circumstances of the project. But some of the principles are interlinked, the implementation process of various projects need to be considered, and some common, the project personnel are easy to understand the problem also needs attention.

BI Modeling Principles

1, modeling the business situation: business needs are the basis, data Warehouse data organization is subject-oriented, rather than report-oriented, is the subject area of business analysis, such as common sales analysis, contract end of the analysis, customer relationship analysis and so on.

2, to ensure the consistency of data: to ensure the correctness and completeness of the logical relationship between data, data Warehouse to achieve the integration of data and data isomorphism, and the relative stability of the data, for the implementation of real-time reading and writing operations.

3, the use of scheduling: Data Warehouse to reflect historical changes and timely and accurate data processing capabilities, so BI modeling incremental update must use scheduling, that is, the fact table for incremental update processing. Consider the actual amount of data before using scheduling, and how often the data is updated. Data volume can be a daily, then the data can be extracted by day, or like the soft business intelligence Finebi, the use of fixed-time incremental updates; it can be a week or more to change. If you have a slow-changing dimension, you need to consider updating the dimension table when scheduling, and update the dimension table before updating the fact table.

4, the balance between demand and reality: according to business needs to provide user-acceptable BI solution, in the BI modeling need to constantly balance user needs and data source facts, not only the designer's own balance, enterprise business people also have to face such a reality.

Problems

1, how to start the model design?

The purpose of BI modeling is to improve management levels, which is the core of the BI project. We must understand the business requirements, business scope and so on, clear the business intelligence expectations and need to analyze which topics. Collaboratively analyze the customer's current level of management, enterprise architecture and operational processes, and the thin weaknesses and key points of the management process to help business people understand their needs.

2, implementation of ignoring the confirmation process

Many project personnel in the confirmation of enterprise needs after the thought can be a bold design implementation, but in the actual process often encountered a variety of problems. The reasons in the early discussion process will always be omitted, some people do not have a deep understanding of the business, resulting in continuous adjustment, project cycle delay. Therefore, in the modeling process, we must constantly confirm the business analysis model, data can be supported. Good Business intelligence BI project implementation, usually fully understand the data extraction object business system, and business personnel to fully communicate, and the leadership repeatedly confirm, avoid the repeat work of enterprise, increase the burden of enterprise.

This paper is reproduced from the sail soft business intelligence finebi:http://www.finebi.com/bi/?p=3362

Bi Modeling Principles and FAQs