| Oracle® Retail Insights Cloud Service Suite Operations Guide Release 18.0.002.5 F25166-01 |
|
 Previous |
 Next |
| Oracle® Retail Insights Cloud Service Suite Operations Guide Release 18.0.002.5 F25166-01 |
|
Previous |
Next |
This chapter outlines the Retail Insights architecture and its components. Review of the architecture is important to understand the data flow and terminology used in the subsequent chapters of this guide.
The following diagram shows the technical architecture of Retail Insights.
ODI resides in Oracle Fusion Middleware suite of products and it is an Extract, Transform, and Load (ETL) tool for extracting and loading data. To use ODI, two database repositories known as the Master Repository and the Work Repository are required.
Refer to the Oracle Retail Insights Installation Guide for more details on installing and configuring ODI.
The Retail Insights staging area is the landing area that has database tables to hold the information from different source applications. All the extracted data from source applications is loaded and transformed into these tables for final loading step.
ODI ETL programs are used for used for extracting data from source tables and for loading into staging tables.
See "Retail Insights Program Overview", "Dimension Data Concepts" and "Fact Data Concepts" for the data flow details.
The Retail Insights data mart area holds the data for all the dimension, fact, aggregate, and supporting tables that are required for data processing. These tables are loaded from Retail Insights staging tables and maintain historical data based on the requirements and configurations. These tables are accessed by Oracle BI EE metadata for reporting and analysis needs. Note that both Retail Insights data mart tables and staging tables exist on the same schema of the database.
The aggregation ETL programs aggregate fact data and move results to Retail Insights aggregation tables. These aggregation tables will be used by Front-end report for better performance. The aggregation programs will aggregate data from temporary tables which are generated in the base loading phase to get better performance.
Choose appropriate aggregate tables based on your business needs as Retail Insights has several levels of aggregates pre-packed. Refer to the Oracle Retail Insights Implementation Guide for additional information on finalizing aggregation strategy.
ODI ETL programs are used for used for loading into data mart tables and also for rolling up/aggregating the data.
See "Retail Insights Program Overview", "Dimension Data Concepts" and "Fact Data Concepts" for the data flow details.
Retail Insights Mining Schema is a separate database schema under the same database with Retail Insights data mart. It is created to contain the starting code for the Retail Insights Data Mining Engine. This schema contains numerous transient tables required for Association Rule Mining (ARM) of sales transaction tables. It also contains some database packages that the Retail Insights weekly batch invokes through the ETL process.
Oracle BI EE is the front end analytical tool used for Retail Insights. The figure below shows the server repository which has the physical mappings and logical modeling of Retail Insights in Oracle BI EE. Users having access to Oracle BI EE metadata can create, edit, and schedule reports based on their roles and requirements.
Oracle BI EE metadata or repository is built on top of Retail Insights data model and includes prebuilt metrics, attributes, hierarchy drills, aggregate navigation, and subject areas.
Web catalog is another component of Oracle BI EE that is used for storing prebuilt interactive dashboards, guided navigation reports and for storing newly created reports.
See the Oracle Retail Insights User Guide for additional information.
