In spy movies, good ‘intel’ is obtained by a clever agents, advanced technologies and lots of pain.
Same applies to organisations: Accurate, insightful and timely business intelligence requires good analysts, much effort and powerful business intelligence applications and technologies to give organisations the edge.
Business intelligence applications typically have the following architecture:
- Transaction data from the organisation’s financial, production, marketing, hr, etc. systems is extracted, transformed and saved into a Data Warehouse (‘DW’) on a regular basis. A DW is a database which structured for efficient storage (called 3rd normalised form) or high speed analysis (called star-schema).
- Where required, data from the DW is pre-processed and stored in data cubes (‘Cubes’). Cubes are conceptually similar to a stack of Excel pivot tables, run for all possible combinations of data. (eg. Sales by Month, Sales by Region, Sales by Employee etc) Cubes are recalculated at various intervals, typically after new data is added to the Data Warehouse.
- Reports and Dashboards are rapidly generated from the Cubes, since most of the data has been pre-calculated and summarised, or directly out of the DW
- Once generated, reports can be displayed on numerous devises (including mobile smart phones), making it easy for employees to monitor company performance whilst on the move. Reports can be exported to a wide variety of formats for further analysis and distribution (eg excel, PDF), whilst ‘self-help’ facilities allow advanced users to write their own reports.
- Security regimes typically control end-users access to data and reports.
The following provides a simple overview of a BI system:
