Data Solutions
Data Warehouse & Business Intelligence

Data Warehouse & Business Intelligence
While operational systems focus on automating internal business processes, they also collect data that can be vital to the prosperity of businesses.

In order to properly leverage it, this data needs to be organized in such a way to serve decision making by either extracting it in separate repository for analysis and reporting, building small data pools, or by implementing business intelligence and reporting tools directly on operational systems.

Selecting the right architecture determines the effectiveness of properly embedding data in the executive decision making processes of an organization.

Centralized Data Warehouse
This architecture is most applicable when large data exists in multiple systems and a heavy analytical requirement exists.
Data is regularly extracted from operational systems and external data sources, properly prepared for storage in the central data warehouse which is modeled specifically to service analytical requirement and flexible data rendering.
In preparing data, consideration must be given to data standardization to avoid reporting gaps, data conflict resolution resulting from the existence of the same data element in multiple systems, and the time needed to collect, prepare and load the data into the central data warehouse repository.

In designing the data model for the warehouse, the typical options of buy versus make are available and careful analysis of the best option could save future agony as model changes proved to be very costly.

Distributed Data Marts
Where specific reporting requirements are needed over a subset of data in the organization, for example credit risk analysis in banks or revenue analysis for telecommunication companies, data marts present a possible alternative.
Data marts contain specific data on a subject matter and are usually used only by a limited set of users.
Data marts can be fed from operational systems directly or from centralized data warehouses.The data model of a mart is most usually custom developed. This requires specialized technical skills in data modeling.
Data marts tend to grow in size as usage becomes more regular. Therefore, careful sizing should be done to ensure mart scalability.

Business Intelligence
Numerous tools are available to facilitate flexible business reporting and analysis, commonly known as business intelligence.
Such tools connect either directly on operational systems or on data warehouses and marts.

BI tools differ in their ability to connect to different data source technologies, the various ways they expose their data service specially in SOA architectures, and in the usability functions they offer to both the normal user and the system administrator.

.