Data Warehouse to BI 1
Agenda Review Preparing the DW for Analysis Microsoft BI Platform Overview Building a Cube in SSAS 2
DW Development Approach: Kimball Methodology DW Project Lifecycle Business requirements Business Requirements Documentation Bus Matrix Design, build and deliver in increments DW Architecture DW Design ETL system Cube, Reports, query tools, … 3
Data Warehouse Project Lifecycle 4 Source: Mundy, Thornthwaite, and Kimball (2006). The Microsoft Data Warehouse Toolkit, Wiley Publishing Inc., Indianapolis, IN.
BI Application Specification, Development Specification Standard Reports Layout-led discovery Enrollments by course and timeframe Average GPA by course and timeframe Analytic Applications Data and Model-led discovery Year-to-year changes in average GPAs by college Students most likely to have good overall GPAs Development (SSAS, Excel) Multidimensional Cube, Data-led discovery (SSRS, Excel) Standard Reports (SSAS, Excel) Mining Model-led discovery 5
Displaying the Cube Cube browser—resembles spreadsheet 6
Displaying the Cube, cont… Pivot table - a multidimensional spreadsheet
OLAP Models ModelDescriptionComparison MOLAPData and aggregates in multidimensional database Best query performance; proprietary ROLAPData and aggregates stored in relational database Worst query performance; leverage existing infrastructure HOLAPHybrid approach; detail data stored in relational database, aggregates stored in multidimensional database Best processing performance; query performance better than ROLAP, not as good as MOLAP 8
Microsoft BI Platform 9 SQL MDX, DMX RDL
Microsoft BI Platform: Development Interfaces SQL Server Database Engine Interface: SSMS SQL Server Integration Services Interface: SSDT SQL Server Analysis Services Interface: SSDT & Excel SQL Server Reporting Services Interface: SSDT MS Office Excel Data Mining Add-In 10
Preparing the Data Warehouse for Analysis Build a multidimensional cube Refine Create KPIs… Design reports Design mining models Deploy cube, reports, models to the server 11
Building an Analysis Services Cube 12 Create: o Data Source o Data Source View o Cube Deploy/Process Cube Use the Cube OLAP Analysis o Reports o Data Mining Models
Internet Example _tutorial.htm 13