Moss L.T., Atre S. Business intelligence roadmap: The complete project lifecycle for decision-support applications
Подождите немного. Документ загружается.
[ Team LiB ]
Risks of Not Performing Step 11
A well-designed and well-tested ETL process is the backbone of a BI decision-support
environment. This step is a very time-consuming step, but without it, you do not
have a BI application. End of story.
[ Team LiB ]
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
351/631
[ Team LiB ]
Bibliography and Additional Reading
Aiken, Peter H. Data Reverse Engineering: Slaying the Legacy Dragon. New York:
McGraw-Hill, 1995.
Beck, Kent. Extreme Programming Explained: Embrace Change. Boston, MA:
Addison-Wesley, 2000.
Cockburn, Alistair. Agile Software Development. Boston, MA: Addison-Wesley, 2002.
Devlin, Barry, Data Warehouse: From Architecture to Implementation. Reading, MA:
Addison-Wesley, 1997.
Hackney, Douglas. Understanding and Implementing Successful Data Marts. Reading,
MA: Addison-Wesley, 1997.
Hetzel, William. The Complete Guide to Software Testing, Second Edition. New York:
John Wiley & Sons, 1993.
Humphrey, Watts S. Winning with Software: An Executive Strategy. Boston, MA:
Addison-Wesley, 2002.
Imhoff, Claudia, Lisa Loftis, and Jonathan G. Geiger. Building the Customer-Centric
Enterprise: Data Warehousing Techniques for Supporting Customer Relationship
Management. New York: John Wiley & Sons, 2001.
Kimball, Ralph, and Richard Merz. The Data Webhouse Toolkit: Building the Web-
Enabled Data Warehouse. New York: John Wiley & Sons, 2000.
Kimball, Ralph, Laura Reeves, Margy Ross, and Warren Thornthwaite. The Data
Warehouse Lifecycle Toolkit: Expert Methods for Designing, Developing, and
Deploying Data Warehouses. New York: John Wiley & Sons, 1998.
Ponniah, Paulraj. Data Warehousing Fundamentals: A Comprehensive Guide for IT
Professionals. New York: John Wiley & Sons, 2001.
[ Team LiB ]
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
352/631
[ Team LiB ]
Chapter Twelve. Step 12: Application
Development
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
353/631
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
354/631
Chapter Overview
This chapter covers the following topics:
Things to consider about access and analysis tools
The advantages of using online analytical processing (OLAP) tools and some of
their popular features
Multidimensional analysis factors, with emphasis on multivariate analysis using
both object dimensions and variable dimensions
Three OLAP architectural layers: presentation services, OLAP services, and
database services
Four common application development environments: prototyping,
development, quality assurance (QA), and production
A short discussion of the Web environment
Brief descriptions of the activities involved in application development, the
deliverables resulting from those activities, and the roles involved
The risks of not performing Step 12
Things to Consider
Prototyping Results
Did we prototype the access and analysis components of the BI
application? If so, what did we learn from the prototype?
What portions of the application prototype can we save? What
portions have to be redesigned?
Did we decide to develop the access and analysis components using
the operational prototype as a rapid and iterative development
method? Will this be the final iteration?
Access and Analysis Tools
What access and analysis tools do we already have in place? Are we
happy with them?
Will the business people use an OLAP tool for multidimensional
analysis?
In what other ways will they analyze the data? Do we need to
acquire another query or reporting tool?
Skills and Training
What skill sets (beginning, advanced, expert) do the business
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
355/631
people have? What type of additional training do they need?
How many business people have already been trained on the access
and analysis tools? How many more need training?
Are they familiar with OLAP and multidimensional reporting
concepts?
Do we need to implement a context-sensitive online help function
for the access and analysis components of the BI application?
Scope and Project Requirements
Are our project requirements still the same? Or has the scope
changed?
How many reports and canned queries do we need to develop?
Which ones?
Have we validated the necessary number of dimensions needed for
those reports and queries?
Do we know how the knowledge workers, business analysts, and
business managers want to drill down and roll up when working with
the data? Are their requirements similar or different?
Web Considerations
Do we have to include a Web portal with this BI application?
Have we developed Web portals before? Have we prototyped one?
What are the additional security requirements for the Web? Do we
have the necessary firewalls and security packages installed?
Technical Considerations
Are the development and production environments configured the
same or are they different?
What are the differences? A different database management system
(DBMS)? A different operating system? Different hardware?
The main reason for a BI decision-support initiative is to provide fast and easy access
to data for business analysis. A high percentage of that access will be by predefined
patterns. A predefined pattern means that data has been precalculated (derived,
aggregated, summarized) and stored in that fashion for faster access. This is the
reason for the high popularity of multidimensional OLAP tools, and it is the hallmark
of BI decision-support applications.
[ Team LiB ]
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
356/631
[ Team LiB ]
Online Analytical Processing Tools
Multidimensional OLAP tools are a major component of the BI decision-support tool
suite. Terms such as OLAP, relational OLAP (ROLAP), multidimensional OLAP
(MOLAP), decision support, multidimensional analysis, and executive information
system (EIS) are all used to describe the explosive growth in the field of data access
and data analysis tools. These terms are frequently associated with expectations of
built-in functionality and ease of use. While much of the literature uses OLAP to
represent all of these terms, each OLAP tool vendor seems to have its own definition
of OLAP. Hence, OLAP tools support only the definitions of their vendors—most of the
time.
A widely accepted definition of OLAP is the following: OLAP refers to online analytical
processing technology that creates new business information through a robust set of
business transformations and calculations executed upon existing data.
Advantages of OLAP Tools
Although BI decision-support applications use conventional reporting and querying
tools as much as multidimensional OLAP tools, the majority of business people seem
to favor OLAP tools because of their additional functionalities. There are two distinct
advantages for business people who use OLAP tools.
The focus in analytical processing is on the data, specifically the
multidimensional aspects of the data that are supported by OLAP tools.
Business objects are represented as dimensions (e.g., product, customer,
department), which are naturally interrelated through functional subject areas
(e.g., sales) and are often hierarchical (e.g., products roll up into product
categories, departments roll up into divisions).
1.
Business analysts navigate through these dimensions by drilling down, rolling
up, or drilling across. They can drill down to access the detailed level of data,
and they can roll up to see the summarized data. They can roll up through the
hierarchy levels of dimensions or to specific characteristics or data elements
(columns) of the dimensions. They can also drill across dimensions to access
the data of interrelated dimensions. In addition, powerful computational
services provide functions such as ranks, averages, return on investment (ROI),
and currency conversions.
2.
The OLAP category of software tools makes many business analysts self-sufficient by
giving them easy and intuitive access to their data for analysis—and getting
developers out of the report-writing business. This is accomplished through the
following tool characteristics:
OLAP tools allow business analysts to combine their data in any order they
desire, at any level of summarization, and over several time periods. Business
analysts can design their queries by clicking on the dimensions and by selecting
the desired data elements for the analysis they need to perform.
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
357/631
Different OLAP tools support a variety of access and analysis needs. They
provide multiple views for data access, from the senior executive's desire to
browse through summarized data to the business analyst's need to perform
complex detailed analysis.
OLAP tools are a very important component of application development in a BI
decision-support environment. While conventional query and reporting tools are used
to describe what is in a database, multidimensional OLAP tools are used to answer
why certain business events are true. For example, an OLAP tool could be used to
prove or disprove a hypothesis that a business analyst has formulated about a
correlation among certain data values. Let us assume that a business analyst makes
the observation that customers with low income and high debt often default on their
loans. The business analyst then concludes that this group of customers should be
considered bad credit risks. In order to verify this conclusion, the business analyst
prepares a query against the BI target databases, for example: "Compare write-off
amounts for products by product type, by territory, by customer, by month, where
the customer income level is below a certain amount to those where the customer
income level is above a certain amount." This type of analysis query would run
against a multidimensional database in which summarized data is stored as
precalculated measures (facts), with one such measure being "write-off amount."
OLAP Tool Features
OLAP tools are popular not only because they make the business analysts more self-
sufficient but also because the tools provide innovative ways to analyze data.
Tools present a multidimensional view of the data, which is intuitive and
familiar to the business people. For example, organizations always like to set
their strategies for increasing revenue by:
- Introducing new products
- Exploring new markets
- Increasing price
These three characteristics are based on increases or decreases in
revenue that can be tracked more easily through a multidimensional view
of the sales data (by product, by region, by customer).
Tools provide summarizations and aggregations of the data at every
dimensional intersection. The terms summarization and aggregation are
commonly used interchangeably. Even though both terms refer to "adding up"
atomic data values and both are used in creating measures or facts, their
precise definitions are not the same.
- Summarization refers to totaling or summing one atomic data value
(vertically) to produce a total or sum of that value, for example, adding
up the annual salaries of all employees to produce the value Total Annual
Salary Amount. (In popular multidimensional database design jargon,
summarization is referred to as aggregation.)
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
358/631
- Aggregation refers to derived data, which is produced from gathering or
adding multiple atomic values (horizontally) to create a new aggregated
value, for example, adding the annual salary, the bonuses, and the dollar
value of an employee's benefits package (health care plan, retirement
plan) to produce the value Employee Compensation Plan Amount.
Tools provide interactive querying and analysis capabilities of the data.
Business analysts can perform "what if" analysis with the help of OLAP tools, for
example, "What if we lowered the price of the product by $5? How much would
our sales volume increase in the state of Alaska?" Business analysts like to run
queries interactively and act upon the query results by changing the values of
some variables and rerunning the query to produce a new result.
Tools support business analysts in designing their own analysis queries, in
creating their own custom members within dimensions, and in creating custom
measures.
- One of the main goals of a BI decision-support environment is to make
the business analysts as self-sufficient as possible. This can be done with
parameterized queries, where business analysts can change their
assumptions (parameters) and rerun the same queries with new
parameters. A prerequisite for effective use of parameterized queries is a
well-documented query library.
- OLAP tools can also give business analysts the ability to create custom
members (also called aggregates) within a dimension (e.g., Hot-Car,
which would be defined as any red convertible) that can SUM, AVG, MAX,
and MIN a group of member values.
- OLAP tools can also provide the ability to create custom measures or
facts (e.g., Percent Female, which would be defined by a formula
provided by a business analyst). These custom measures can then be
picked as a new measure from a drop-down menu for a fact table.
Tools support drill-down, roll-up, and drill-across features of
multidimensional analysis. For example, a business analyst who wants to find a
way to lower the cost of manufactured goods could drill down into the actual
detailed costs of purchased raw materials. He or she could also summarize
these costs by rolling up the raw materials into predefined categories. Then he
or she could drill across to another table to include the production costs of the
manufactured goods.
Tools offer analytical modeling capabilities useful to business people. To
expand on the previous example, lowering the cost of manufactured goods
could also be accomplished by reducing the working capital so that the
borrowing costs are lower. Analytical modeling techniques provided by the
OLAP tools could be used to find the optimum amount of working capital.
Tools support functional models for trend analysis and forecasting. OLAP
trend analysis functionality could be used to analyze past data and make
predictions about the future.
Tools display data in charts and graphs that offer quick visual summaries.
The saying "A picture is worth a thousand words" has never been so true as
when analyzing vast amounts of data from BI target databases. Visually
appealing, understandable, and useful charts and graphs are important
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
359/631
components of every OLAP tool.
There is a diversity of OLAP tools based on different architectures and features. It is
important to know the access and analysis requirements of BI applications in order
to make an informed decision about purchasing the right OLAP tools.
[ Team LiB ]
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
360/631