设计数据仓库的注意事项(Part I)

原创 2001年09月14日 21:31:00

Introduction

Identifying and understanding the business and operational requirements that drive any data warehousing project are essential to the project抯 successful completion. You cannot meet your clients?needs if you have not assessed and analyzed what those needs are and how you can address them.

Who Knows?

The most important part of the business requirements process is finding out who knows what the requirements should be. In most cases, the answer to "Who knows what we need to know?" is that nobody knows it all, but everyone knows a part.

And What Do They Know?

The answer to this question is they know far more than they imagine. In most cases your sources do not know they have valuable information and to what extent they are the only possible source for vital information that will help insure the project抯 success.

The Biological Knowledge Base

Systems are designed, created, maintained and operated by people who have first hand system information. This group of experts is collectively identified as the biological knowledge base that you must tap effectively in order to extract the information you require.

Play by the Rules

Your organization抯 business rules (functional requirements) and operational characteristics (non-functional requirements) are the essence of the design requirements for your project. They must be carefully cataloged and analyzed to insure internal consistency. When inconsistencies are found, you need to identify the cause of the inconsistencies (you will find that in some cases the inconsistencies are perfectly valid and reasonable) and determine the proper method(s) for adjusting the project.

How It Has to Work and Why

For each component of your project, you will need to identify the operational and business requirements. For example, there will be times when the need for 24X7 availability will supercede the need for total accuracy. Situations such as this must be taken into consideration and included in your development plan.

Cultural/Political Constraints

Cultural and political constraints will have a serious impact on your success, so they must be addressed while collecting the business requirements that define the project. Even in those cases where the political environment is opposed to the needs of the project team, you can use political relationships to increase the visibility and acceptance of your project, but only if you have a clear understanding of the dynamics involved.

Assessing/Addressing Risks

Every project faces risks, some are minor, some severe. The successful project is based on identifying what those risks are, how likely they are to occur, how serious they are and what actions are best to prevent the issues from arising. Identify the tactics to be used in dealing with the issues when they do arise, because, despite your best efforts, some will still occur.

Mine the Biological Knowledge Base

The biological knowledge base is the complete body of human knowledge and intelligence available. The readily accessible biological knowledge base primarily consists of: designers and developers, knowledge workers, end users, management, system staff, security staff and physical maintenance staff.

Everything you learn from the biological knowledge base should become a part of the business requirements for your project and be fully documented to serve as valuable material for subsequent projects.

Designers/Developers

Design and development staff members are those who are or have been responsible for creating and maintaining existing systems and are tasked with integrating future systems. You will need to gather business requirements and functional constraint information from architects (business, data and system), database administrators, database analysts and programmer analysts. Each will have their own perspective on your project and on the information needed in order to make it successful.

Effectively interviewing the designers and developers will provide you with background information that will give your project a head start toward a successful completion. Their knowledge will provide a foundation upon which to build the final business requirements that your business partners identify.

Architects

Architects create the designs for the environments in which applications and systems function. They are seldom directly involved in building the applications, systems or databases, but provide the blueprints from which such structures are created.

Architects can detail information contained in the design documentation as well as the enterprise standards, logical limitations and business functionality. Architects also begin gathering the data requirements for the project.

DBAs

Database administrators and analysts are experts when it comes to creating, maintaining and optimizing the data structure behind information management applications. These DBAs are an invaluable resource.

DBAs can provide your team with nearly everything it needs to know about database management systems, structure, security, resource requirements (hardware, software and personnel), enterprise standards, physical limitations and how others have integrated business rules into their database implementations.

Programmer Analysts

Programmer analysts are the ones who create the code that makes applications work. Their job is to take a concept and turn it into a functioning system that provides the desired capabilities.

These specialists should be interviewed with the intention of deriving information on operation limitations in their applications, structural standards that constrained the design and development of the applications, how they enforce the business rules that defined their projects and a listing of the design requirements from which they worked.

Knowledge Workers (Subject Matter Experts)

Knowledge workers are those who work with the base information for your project. They consist of business analysts, actuaries, accountants and business consultants, as well as marketing, sales, clerical and production.

Business Analysts

Business analysts are your best resource for understanding the true business meaning for a particular data element or set and for identifying the logic on which the application is based.

They will provide you with information on which business questions need to be answered and where to find the data that will help to answer them. They will also provide you with the specifications on functional reporting requirements ?the information they need to have in order to develop the reports upon which decisions are based.

Actuaries

Actuaries are statistical analysis specialists who determine the likelihood that something will happen a specific number of times within an identified group during a given time period. They can also be called risk analysts because they job identify risks the business faces.

They provide your team with information on the data elements they require in order to perform the statistical analysis upon which their predictions are dependent. They can also identify many of the derived and dependent measures that will assist them in preparing analyses as needs and demographics change.

Accountants

Accountants are the true numbers crunchers of the organization. They are the ones who determine to what extent the organization is profitable. They make sure the books balance. They tend to be detail oriented to the exclusion of nearly every other consideration.

You will get specific information regarding the data that feeds the profit/loss analyses they use in building financial reports and budget management reports. They will also provide detail on the derived and calculated measures from which they build these reports.

Business Consultants

Business consultants, whether internal or external, specialize in identifying ways to improve business functionality and preparing plans for implementing the new processes. In order to accomplish their tasks, they must receive clear and reliable information regarding the current and historical views of the business as well as developing trends.

The consultants will identify the business questions they need answers to in order to identify the current state, which they base their trend analyses on. In many cases, the most important pieces of information are derived from profit/loss reports.

Marketing Staff

The marketing team is primarily interested in identifying potential customers, learning what products or services they want and determining how your organization can fill all or part of that need. In order to perform their tasks most effectively, they need to have sales, production and customer support data.

The marketing team will be able to provide you with information on the following topics: client segmentation, test marketing requirements, marketing program effectiveness, customer relationship management and value chain management.

Sales Staff

The sales staff is responsible for actually turning a marketing plan into financial transactions. They need to judge the effectiveness of sales strategies and tactics on a near real- time basis in order to adjust to market conditions.

Your sales professionals will identify the information they need in order to support both sales and production reporting. They will need this information on a periodic basis (sales for a specific date, week or month) as well as on a period to date basis (production so far this year or month). They will also need reports detailing projected production, inventory levels and sales in queue.

Clerical Staff

The clerical staff is primarily responsible for all data entry operations. They are usually hourly employees who do not have a highly developed technical skill set. They often overlooked in determining the design requirements for a business intelligence data store. However, they can be a font of knowledge regarding productivity reporting and data quality issues.

When interviewing these team members, focus on the data issues that effect their productivity. They will be able to help you identify the elements which are essential to developing a comprehensive data quality program as well as staff productivity reporting.

Production Staff

Your production staff is most intimately involved in developing, producing and delivering your organizations? products and services. The information that most directly affects production includes: pending orders, inventory (current and projected), changes in production requirements (especially special marketing) and personal productivity.

Production is another group that is frequently underutilized in determining data warehouse requirements. They can provide vital information on the data elements such as inventory control and order processing reporting systems. They will also be your best resources for validating that the productivity requirements you have identified are in fact those that have the highest value for the organization.

Other

Depending upon the structure of your organization and the business you are in, you may have several other categories of knowledge workers you will need to interview. Some examples include: legal, political action teams, environmental impact specialists, medical practitioners and product design professionals. In each case, take the time to get to know the job as well as the person doing it, so you can understand the context in which the required information is used.

数据仓库的架构与设计

公司之前的数据都是直接传到Hdfs上进行操作,没有一个数据仓库,趁着最近空出几台服务器,搭了个简陋的数据仓库,这里记录一下数据仓库的一些知识。涉及的主要内容有: 什么是数据仓库? 数据仓...
  • Trigl
  • Trigl
  • 2017年04月01日 17:52
  • 6497

设计并实现数据仓库ETL过程(IBM讲座)

本文是关于计划、设计和实现基本数据仓库解决方案的系列文章的第 3 部分,将设计和实现仓库 ETL 过程,并了解仓库的性能和安全问题。 请多多支持作者作品RKWeb1.1 asp.net开发模板! ...
  • pianzif
  • pianzif
  • 2014年07月14日 12:56
  • 2068

数据仓库主题设计及元数据设计

数据仓库主题设计及元数据设计
  • zhangzhongzhong
  • zhangzhongzhong
  • 2016年04月15日 15:06
  • 7589

数据仓库应用(一):数据仓库模型设计

关联文章: 数据仓库应用(二):数据抽取、转换、加载(ETL) 数据仓库应用(三): SQL Server 2005的数据仓库应用–联机分析OLAP一、问题背景 某电子商务网站主要销售电子类...
  • lxf_44944
  • lxf_44944
  • 2015年02月14日 23:02
  • 2483

2014年总结之dm组数据仓库设计总结

来搜狐移动视频dm组已经快一年半了,今天是2014年12月31号,是2014的最后一天,想想写点东西。写点什么尼,咱们就说说数据仓库dm组设计的一些实践把。 我这里也感谢我得领导给我的机会和舞台。ok...
  • map_lixiupeng
  • map_lixiupeng
  • 2014年12月31日 16:03
  • 3524

数据仓库的模型设计

分类: 数据仓库与数据挖掘 Technorati 标签: 数据仓库,模型设计 数据仓库的模型设计 A. 数据建模方法论 数据仓库模型设计遵循“自顶向下、逐步求精”的设计...
  • zhaojike
  • zhaojike
  • 2017年04月05日 00:12
  • 558

Hadoop学习笔记(10)-简述分布式数据仓库Hive原理

0.什么是数据仓库数据仓库,英文名称为Data Warehouse,可简写为DW或DWH。数据仓库,是为企业所有级别的决策制定过程,提供所有类型数据支持的战略集合。它是单个数据存储,出于分析性报告和决...
  • xummgg
  • xummgg
  • 2016年10月04日 23:32
  • 982

数据仓库(八):数据仓库中数据库设计的基本模型对比--Inmon

注:本文参考Inmon的著作,Inmon支持关系模型,因此下面的观点基本都是支持关系模型的。关于多维模型的内容可参考Kimball的著作《数据仓库工具箱:维度建模的完全指南》等。 广泛采用的数据库设...
  • kingzone_2008
  • kingzone_2008
  • 2013年05月20日 15:42
  • 6999

数据仓库构建步骤

概述: 1. 数据仓库是面向主题的、集成的、不可更新的、随时间的变化而不断变化的这些特点决定了数据仓库的系统设计不能采用同开发传统的OLTP数据库一样的设计方法 2. 需求不明确  数据仓库...
  • u011544778
  • u011544778
  • 2016年03月16日 18:38
  • 689

数据仓库的模型设计

数据仓库的模型设计 A. 数据建模方法论 数据仓库模型设计遵循“自顶向下、逐步求精”的设计原则。 模型设计分为三个阶段: 1,概念模型 对业务的范围和使用,从高度上进行抽象概...
  • xiaolang85
  • xiaolang85
  • 2016年08月03日 16:27
  • 984
内容举报
返回顶部
收藏助手
不良信息举报
您举报文章:设计数据仓库的注意事项(Part I)
举报原因:
原因补充:

(最多只允许输入30个字)