Business intelligence technology comprises of a collection of decision supported technologies designed to enable knowledgeable workers in a business such as managers, analysts and executives to make faster and more informed decisions. Over the last two decades, there has been explosive growth in the number of products and services offered, as well as in the adoption of business intelligence technologies by industry. This growth has been due to the declining cost of obtaining and storing large data amounts that arise from retail, customer banking transactions, as well as in email, e-businesses, query logs for websites, inventory tracking RFID tags, product reviews and blogs.
Today enterprises are experimenting with and using more sophisticated methods for data analysis in making decisions and delivering new features such as personalised services and offers to their clients and successful businesses are turning to business intelligence technology to accomplish this. For instance, BI technology may be employed in manufacturing for customer support and order shipment, in retail to profile users when targeting grocery coupons during checkout, in financial services for detection of fraud and analysis of claims, in transportation fleet management, in telecommunications for the identification of reasons for customer churn, in health care for analysis of outcomes and in utilities for the analysis of power usage.
Tasks are performed over data that are sourced typically from multiple operational databases across various departments within an organisation, and external vendors. The varying sources result in data that is of equally varying quality, using inconsistent codes, formats and representations that have to be reconciled. In order for BI to be performed without problems of cleansing, integration, and data standardisation, you will require efficient loading of data. Furthermore, these tasks typically need to be incrementally performed as new data is received, for instance sales data from last month. As such, scalable efficient data loading, as well as refresh capabilities are essential for every business. Back-end technologies for the preparation of BI data are collectively known as Extract-Transform-Load or ETL tools. Specialised engines known as Complex Event Processing engines are available to provide support for tasks in near real time, thereby enabling businesses to make decisions based on the operational data itself.
The data over which business intelligence tasks are to be performed is normally loaded into a repository referred to as the data warehouse which is managed by one or more data warehouse servers. The relational database management systems or RDBMS is a popular engine for the storage and querying of warehouse data. The last two decades have witnessed the development of various data structures, optimisations and techniques for query processing to enable the execution of complex SQL queries over enormous volumes of data – which is an essential requirement for BI. An example of such an ad hoc SQL query would be to identify customers who during the last quarter have placed orders in amounts exceeding the average order amount by at least 50 per cent. A large data warehouse will typically deploy parallel RDBMS engines such that the SQL queries can be executed with low latency, over large data volumes.
About the Author
Sarah is a Search Consultant at a Digital Marketing Agency: http://www.fdcstudio.co.uk/