Moss L.T., Atre S. Business intelligence roadmap: The complete project lifecycle for decision-support applications

Подождите немного. Документ загружается.

- Quality

- Cost

- Additional skills

or resources

required

Review the impact of

requested changes with

the business sponsor and

the business

representative

Revise the application

requirements document

to include approved

changes

Review lessons learned

with the entire project

core team and in

particular with the ETL

step core team

Use prototype

demonstrations to

promote the BI

application

Perform the next

prototype iteration, if

applicable

7. Meta Data Repository

Analysis

1. Analyze the meta data

repository requirements

Review the technical

infrastructure

assessment report for

requirements to license

(buy), build, or enhance

a meta data repository

Perform a cost-benefit

analysis for licensing

versus building a meta

data repository

Make the decision to

license or build a meta

data repository

Review the nontechnical

infrastructure

assessment report for

new meta data

requirements (business

and technical)

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

521/631

Determine the scope of

the meta data repository

deliverables

Prioritize the meta data

repository deliverables,

indicating whether the

meta data components

are mandatory,

important, or optional

Update the application

requirements document

to reflect any changes

2. Analyze the interface

requirements for the meta

data repository

Analyze the meta data

sources from which the

meta data components

will be extracted:

- Word processing

files and

spreadsheets

- DBMS

dictionaries

- CASE, ETL, OLAP

tools

- Report writers

and query tools

- Data mining tool

Determine what import

and export features are

available in these tools

as well as in the meta

data repository product

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

522/631

3. Analyze the meta data

repository access and

reporting requirements

Review the original meta

data repository access

and reporting

requirements

Review the meta data

security requirements

Identify the access

interface media for

displaying meta data ad

hoc query results (for

example, PDF, HTML)

Analyze the feasibility of

a context-sensitive help

function

Determine what reports

should be produced from

the meta data repository

4. Create the logical meta

model

Create business meta

data entities

Create technical meta

data entities

Determine the

relationships between the

meta data entities

Create attributes for

business and technical

meta data entities

Draw an entity-

relationship diagram

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

523/631

5. Create the meta-meta

data

Describe all meta data

entities with:

- Name

- Definition

- Relationships

- Security

- Physical location

- Timeliness

- Volume

Describe all meta data

attributes with:

- Name

- Definition

- Type and length

- Domain (content)

- Security

- Ownership

Define the business rules

for meta data entities,

attributes, and

relationships

8. Database Design

1. Review the data access

requirements

Review the data-

cleansing specifications

Review the prototyping

results with the

application lead

developer

Review detailed access

and analysis

requirements with the

application lead

developer and the

subject matter expert or

business representative:

- Reporting

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

524/631

requirements

- Querying

requirements

- Ad hoc querying

requirements

Review data security

requirements

Determine projected data

volumes and growth

factors

Determine the projected

number of concurrent

database usages

Determine the location of

business people

Determine the frequency

of report and query

executions

Determine the peak and

seasonal reporting

periods

Determine platform

limitations

Determine tool

limitations (ETL, OLAP,

report writers)

2. Determine the

aggregation and

summarization

requirements

Review the measures

(facts) used by the

prototype (how they

were derived through

aggregation and

summarization)

Review the dimensions

used by the prototype

Review the drill-down

and roll-up functions of

the prototype

Review common

reporting patterns

among existing reports

and among business

people from various

departments

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

525/631

Determine the most

frequently used reporting

dimensions

Review the logical data

model with the data

administrator

Determine the level of

detail (granularity)

needed

Determine how the

detailed data will be

accessed (drill-down or

ad hoc)

Determine how many

business relationships

(entity relationships)

among detailed data will

be needed, if any

3. Design the BI target

databases

Determine the

appropriate database

design schemas:

- Multidimensional

star schema

- Multidimensional

snowflake schema

- Entity-

relationship-based

relational schema

- Hybrid design

schema (mixture)

Create the physical data

models (database design

diagrams)

Create the technical meta

data for the physical data

models (for example,

names and definitions for

tables, columns, keys,

indices)

Map the physical data

models to the logical

data model

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

526/631

4. Design the physical

database structures

Determine how to cluster

the tables

Determine the placement

of datasets

Determine how to stripe

disks

Determine how to

partition the tables

across multiple disks

Determine how much

free space to choose

Determine how much

buffer space to declare

Determine how large to

set the blocksize

Determine the most

appropriate indexing

strategy

Determine whether

referential integrity will

be enforced by the DBMS

or by the ETL programs

5. Build the BI target

databases

Create the data definition

language (DDL) defining:

- Storage groups

- Databases

- Partitions

- Tablespaces

- Tables

- Columns

- Primary keys

- Foreign keys

- Indices

Create the data control

language (DCL) to:

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

527/631

- Define

parameters for the

security SYSTABLE

- Set up group IDs

- Grant CRUD

(create, read,

update, delete)

authority to the

group IDs

- Assign

developers,

business analysts,

and programs to

the appropriate

group IDs

Run the DDL to create

the physical database

structures

Run the DCL to grant

CRUD authority to the

physical database

structures

Build the indices

6. Develop database

maintenance procedures

Define database

maintenance activities

for:

- Database

backups (full

backups and

incremental

backups)

- Disaster recovery

procedures

- Reorganization

procedures for

fragmented tables

Define the frequency of

and procedure for

performance monitoring

activities

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

528/631

7. Prepare to monitor and

tune the database designs

Note

This is an

ongoing post-

implementation

activity.

Plan to monitor the

performance of ETL

loads, reports, and

queries at runtime

Plan to use a

performance-monitoring

utility to diagnose

performance degradation

Plan to refine the

database design schemas

Plan to add additional

indices, if necessary

8. Prepare to monitor and

tune the query designs

Note

This is an

ongoing post-

implementation

activity.

Plan to review and

streamline all SQL calls

in ETL programs and

application programs

Plan to write pass-

through queries for OLAP

tools, if necessary

Plan to utilize parallel

query execution

Extract/Transform/Load

Design

1. Create the source-to-

target mapping document

Review the record

layouts for the source

files

Review the data

description blocks for the

source databases

Review the data-

cleansing specifications

for source data elements

with the data quality

analyst, the subject

matter expert, and the

business representative

Create a matrix for all

target tables and target

columns

List all applicable source

files and source

databases for every

target table

List all relevant source

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

529/631

data elements for every

target column

List data type and length

for every target column

List data type and length

for every source data

element

Write transformation

specifications for

populating the columns:

- Combine data

content from

multiple sources (if

needed)

- Split data content

from one data

element across

multiple columns if

source data was

used for multiple

purposes

- Include

aggregation and

summarization

algorithms

- Include data-

cleansing

specifications for

each column

- Include logic for

checking referential

integrity (if not

performed by the

DBMS)

- Include logic for

error messages

and record

rejection counts

- Include logic for

reconciliation totals

(record counts,

domain counts,

amount counts)

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

530/631