How does BI work?

The BI concept is relatively simply in theory, but sometimes harder to deploy in reality!

Data from 'source systems' (usually the organisation's operational applications, databases or spreadsheets) are periodically (often daily) extracted, processed/cleaned and stored into a data warehouse (this process is known as Extract, Transform and Load, or ETL).

The 'transform' part of ETL includes ensuring that data from the various sources can be used seamlessly. For example, if one individual appears on the HR system as an employee, the research system as a project member, and the education system as a postgrad research student, ETL will ensure that the individual's 3 unique identifiers are linked in a meaningful way to allow a single BI report to include data from all three systems.

BI then uses a variety of tools to retrieve data from the warehouse and to present it to users.  A key characteristic of this, not unique to BI but certainly more prevalent in BI culture than in much traditional MI, is that reports are often interactive with their users, which may include any or all of the following:

  • proactive (for example alerting a manager if a preset trigger is reached, such as the numbers of admissions offers for a particular programme reaching the target number)),
  • responsive (for example allowing users to select tailor-made and dynamic reports, rather than reports being static and centrally-disseminated)
  • contextual (for example, automatically identifying a user’s profile and adjusting the content or emphasis of a report to reflect their role in the organisation or their previous usage. 

 The page 'Our Current Projects' describes how we are implementing BI at Exeter