You are on page 1of 15

<Insert Picture Here>

Best Practices for Data Warehousing

Agenda Best Practices for DW-BI


Best Practices in Data Modeling Best Practices in ETL Best Practices in Reporting
<Insert Picture Here>

Best Practices in Data Modeling

Support for Cross-Application Analysis

Dim Table Dimension Dim Table Tables

Products Dimension

Dim Table Dimension Dim Table Tables

Sales Orders Fact Table

Purchase Orders Fact Table Time Dimension

Oracle Order Management & Fulfillment Analytics

Supply Chain Analytics

Q. How many of my top customers bought products from my worst suppliers?

Fundamental requirement that dimensions be common (conformed)

Integrated Enterprise Analytics Data Model

Customers

Procurement

Distribution

Operations

Customers

HR / Workforce

Suppliers

Marketing

Service

Finance

Sales

Suppliers

Customers

Suppliers

Features:
Conformed dimensions Transaction data stored in most granular fashion Tracks full history of changes Prebuilt and extensible Built for speed

Benefits:
Enterprise-wide business analysis (across entire value chain) Access summary metrics or drill to lowest level of detail Accurate historical representations

The Result
From this :

To this :

- Fewer, larger database tables rather than many smaller ones - Same piece of data appearing in several locations - Reduces need for join paths - Structure is denormalized for performance

Best Practices in ETL

ETL Architecture Best Practice


Presentation Dashboards by Role Reports, Analysis / Analytic Workflows

Administration

Physical Map Data Warehouse / Data Model Direct Access to Source Data Load Process Staging Area ETL

Staging Tables

DAC

Source Dependent Extract


SQL SQL SQL
Special Connect Special Connect

App Layer

Siebel OLTP

Oracle

App Layer

Other

Oracle SAP R/3

Siebel

PSFT

EDW Other

PeopleSoft SAP 8

Federated Data Sources

Extract

ABAP

Extraction Process

SQL

Transform

Logical Model / Subject Areas

BI Server

Metadata

Metrics / KPIs

Source Independent Layer

Load

Data Warehouse

Use of a ETL platform

Limited programming GUI interface Re-usable components Easy data lineage tracking (where did data come from?) Pseudo-documentation fast ramp-up for new resources Can build, test & implement the data flows more quickly

ETL Framework Best Practices

Generates surrogate key


Does lookups for descriptions of code fields

Does data driven updates inserts for new rows, updates for old rows
Reject Capture

Keep track of effective dates and maintain history as required


Handles Deletes

10

Best Practices in Reporting

11

The Semantic Layer


Multi-layered Abstraction
Dashboards by Role Reports, Analysis / Analytic Workflows BI Presentation Services

Administration

Logical Model / Subject Areas


Physical Map Data Warehouse / Data Model Direct Access to Source Data Load Process Staging Area

BI Server

Metadata

Metrics / KPIs

Separation of physical, logical and presentation layers Logical modeling builds upon complex physical data structures Logical model independent of physical data sources, i.e. same logical model can be remapped quickly to another data source

Metrics / KPIs Aggregate navigation Prebuilt hierarchy drills and cross dimensional drills

ETL

Extraction Process

Oracle SAP R/3

Siebel

PSFT

EDW Other

Federated Data Sources

DAC

12

Object Security
What parts of the application can you see? Business Logic Object Security
Presentation Layer

Object Security

Semantic Object Layer


Physical Layer

Controls access to Subject Areas, Tables and Columns in Presentation Layer

Web Object Security


Limits access to Dashboards, Reports and Web Folders

DW -BI Architecture Best Practice


Additional dashboards and reports, guided and conditional navigations, iBots, etc. Additional derived metrics, custom drill paths, exposing extensions in physical, logical and presentation layer, etc. Extension of DW Schema for extension columns, additional tables, external sources, aggregates, indices, etc. Extension of ETL for extension columns, descriptive flexfields, additional tables, external sources, etc.

Dashboards & Reports

Easy

Semantic Metadata Layer

Moderate

DW Schema

Intermediate

ETL

Involved

Degree of Customization

Level of Effort
14

For any sales queries, please contact Oracleindia_in@oracle.com

15

You might also like