Upload
cordelia-flynn
View
216
Download
1
Embed Size (px)
Citation preview
Introduction to Building a BI Solution
권오주
OLAPForumwww.olapforum.com
Agenda
Introduction to Building a BI Solution
Defining Data Warehouse Structures
Populating the Data Warehouse
Creating Simple and Parallel Data Loads
Copying, Managing, and Transforming Data
Storing, Managing, and Executing Packages
Agenda
Building Cubes
Understanding Analysis Services Architecture
Designing OLAP Dimensions
Using Measures
Implementing Calculations Using MDX
Applying Virtual Cubes
Case Study
Agenda
Designing Aggregations
Processing Cubes and Dimensions
Optimization and Performance Tuning
Advanced Analytical Features and Security
Using Office for Reporting and Analysis
Jumpstarting BI Solutions with SSABI
Understanding Business Intelligence (BI)
The Primary Goal of BI Is to Impact the Bottom Line
Strategic decisions protect and enhance competitive advantage
Tactical decisions manage and measure specific operations or employee behavior
BI Requires the Integration of Several Components
Data that companies collect in its daily operations
Technology which collects and organizes data
People who can analyze data and make effective decisions
Components of Multidimensional Analysis
Measures
Numeric values that are of interest to business analysis
Base measures, Key Performance Indicators (KPIs), and benchmark metrics
Dimensions
Categorical view of measures
All members of a dimension belong together as a group
Single-Dimension View
Each view appears homogenous
Product TotalApples 8000Cherries 8000Grapes 8000Melons 8000Grand Total 32000
Quarter TotalQtr 1 16000Qtr 2 16000Grand Total 32000
Region TotalAtlanta 8000Chicago 8000Denver 8000Detroit 8000Grand Total 32000
Multi-Dimension View
Interaction reveals variations
Units RegionProduct Quarter Atlanta Chicago Denver Detroit Grand TotalApples Qtr 1 4000 4000
Qtr 2 2500 1500 4000Apples Total 4000 2500 1500 8000Cherries Qtr 1 1000 3000 4000
Qtr 2 2000 2000 4000Cherries Total 1000 3000 2000 2000 8000Grapes Qtr 1 1500 2500 4000
Qtr 2 1000 3000 4000Grapes Total 1000 3000 1500 2500 8000Melons Qtr 1 2000 2000 4000
Qtr 2 2000 2000 4000Melons Total 2000 2000 2000 2000 8000Grand Total 8000 8000 8000 8000 32000
Benefits of Multidimensional Analysis
Useful approach for viewing information
Allows for flexible business analysis
Ability to analyze measures simultaneously categorized across many dimensions
Especially powerful when measures and dimensions include combined data from multiple data sources
OLTP Data Sources
OLTP System Characteristics
Processing real-time transactions of a business
Containing data structures optimized for entries and edits
Providing limited decision support capabilities
OLTP System Examples
Order tracking
Customer service
Service-based sales Banking functions
Silos of Data
CallCenter
CallCenter
MarketingCampaign
Mgmt
MarketingCampaign
Mgmt
CRM and eCRM
CRM and eCRM
Internet
Internet
Financial/ Accountin
g
Financial/ Accountin
g Procurement
Procurement
HRHR
Inventory
Inventory
Data Warehouse
Data Warehouse System Characteristics
Presents Data for Business Analysis Processes
Provides a Consistent Historical Data Store
Stores Data in Structures that Are Optimized for Extraction and Querying
Integrates Data from Heterogeneous Source Systems
Combines Validated Source Data
Organizes Data into Non-Volatile, Subject-Specific Groups
Data Warehouse System Components
Data Warehouse
Data Access
End User Data AccessData
Sources StagingArea
Data Marts
Data Extract, Transform, and Load
The Microsoft Data Warehousing Framework
Microsoft SQL Server 2000
Relational Database Management Systems
Contain and manage large quantities of data
Foundation of a data warehouse
SQL Server 2000 Roles
Online Transaction Processing System
Data Staging
Data Warehouse
Data Mart
Data Transformation Services
Extract, Transform, Load, and Management (ETLM) Tools
Extract data from heterogeneous source systems
Transform source data to load into a destination
Data Transformation Services
Copies and transforms data from a variety of sources
Creates reusable transformations and functions
Automates data loads
Analysis Services 2000
Online Analytical Processing (OLAP) Databases
Provide an intuitive, multidimensional view of data Provide fast data retrieval Robust calculation engines
Analysis Services 2000
Creates multidimensional cubes Optimizes aggregations to provide rapid response Supports multidimensional expressions (MDX) to retrieve and
manipulate multidimensional data Includes PivotTable service, an OLE db-compliant provider, for
reporting applications
Extensible Markup Language for Analysis
XML/A is a Data Access Protocol Extending BI to Microsoft .NET Platform
Supports Exchange of Analytical Data Between Clients and Servers
Available on any device or platform
Using any programming language
End User Data Access
End User Applications
Data Access and Distribution Mechanisms
Ad-hoc Query Tools
Report Writers
Modeling Applications
Portals and Dashboards
Excel for Business Intelligence
Excel Provides a Familiar Interface for Data Analysis
OLAP PivotTables and PivotCharts Allow Access to Large Data Sets
Offline Cube Files Allow Analysis When Disconnected from the Network
Office Web Components for Business Intelligence
OWC Delivers PivotTable and PivotChart Functionality to Web
OWC Facilitates Flexible, Customizable Solutions
Data Analyzer for Business Intelligence
Data Analyzer Adds Rich Visualization and Analysis Capabilities
Integration with Excel Facilitates Exploration Before In-Depth Analysis
Full Support for OLAP features in Analysis Services
SharePoint Portal Server
Organizing Documents
Finding Documents
Implementing Approval Processes
Ensuring Document Security
Searching for Documents Collaboration and Update Notification Providing Scalability at the Enterprise
Level
Microsoft SQL Server Accelerator for BI
Extension of Microsoft BI Platform Used to Build Customizable BI Solution
SQL Server Accelerator for BI Components
Analytics Builder Workbook to configure data model
Analytics Builder utility to create an analytical application based on the configured data model
Templates for business analysis with front-end tools
Imagine the Possibilities