What's Business Intelligence ? Decision Support System

" We are just beginning to understand how to use information as a tool " Peter Drucker

DSS. The Data Warehouse project. Business Intelligence's Software

Business Intelligence proposes to use the data transiting through the enterprise database system, as informations to be exploited with a decisional purpose.
Technically and practically, the Business Intelligence is composed of a family of computer systems and softwares to guarantee the mechanism of the decisional-informations making chain (Decision Support Systems DSS). For a best understanding the elements composing the decisional chain are presented in four categories each one with a specific function.
Gathering
To clean and consolidate the data. To remove the data from the production systems and to adapt them for a decisional use.
Stocking
To concentrate the structured and processed data so that they are available for a decisional use.
Dispatching
Or to ease the accessibility to the data according to the functions and the kinds of use.
Exploiting
Or how to find the best way to help the user so that he can remove the information's substance from the data stocked in that purpose.

1 Collecting : ETL tools (Extract Transform and Load)

Collecting, cleaning and consolidating the data of the Extended Enterprise Company.
The data collect is a function fulfilled by a family of tools known as ETL for Extract Transform Load. The company information system wasn't built in one day. The main part of the company information systems are heterogeneous. Although the standardization of exchanges between various computer tools is striding, the disparity of the data formats is still a reality. It's the main technological gap to extended data exchanges.
Before being useable, the data should be formated, cleaned and consolidated. The ETL tools allow to automate these processings and to manage the data flows supplying the storage bases : Data Warehouse or Datamart.
Note that the Meta data system like Common Warehouse Metamodel (CWM) keeps accurate characteristics about the data, in other words data about data.

2 Stocking : Data Warehouse and Datamart

The enterprise database system can't be used for a decisional exploitation. The data are not ready for this use in addition the decisional requests are costly for machine resources.
The data, first cleaned and consolidated, are stocked in a specialized base : Data Warehouse or Data mart.
Data mart is a Data Warehouse's smaller version. Datamart is directed towards the subject or the theme and can be, for example, used for CRM (Custom Relationship Management) or Data Mining applications.
Data Warehouse or Datamart are supplied by the ETL tool. Note, that the Data Warehouse project is quite particular. It is preferable to consider it as a process. Data Warehouse is indeed in perpetual evolution.

3 Distributing information : EIP Enterprise Information Portal

The empowerment and the increase of the decision making points alter fundamentally the information management. The information will be perceived as a flow and not as a stock unit. In order to revitalize the global reactivity, the information will be widely distributed to the partners set. The EIP Enterprise Information Portal, fulfill this main role.

4 Exploiting : dashboards, KPI, OLAP, data mining, ...

Once data are stocked, cleaned, consolidated and accessible, they are useable. According to the needs, various kinds of extraction and exploitation tools can be considered.
- To analyze the data with OLAP tools for multidimensional analysis.
- To look for little visible correlations with Data Mining.
- To assist the decision-maker in situation, with dashboards exposing Key Performance Indicator (KPI).
- To communicate the performance with Reporting.

Main Open Source products
Pentaho, Spago, BEE and also Marvel-IT are global projects covering the whole decisional chain.

---

Copyright : Alain FERNANDEZ ©1998-2009- Tous droits réservés

---