Shahla Butler would like to thank her colleague, Tricia Spencer, for contributing this month's column.

Like many analysts in data warehousing, I subscribe to the need for multiple tools to access and gain value from my data warehouse. Some of these tools were introduced by my colleague, Shahla Butler, in last month's column. In the architectural framework she described, the business intelligence component identifies retrieval, analysis and discovery technologies for accessing the data warehouse.

Retrieval technologies are typically report and query tools for when the user knows exactly what data is needed. These tools may often format that information in a way that is easy to communicate to a wide audience. OLAP tools go a step beyond with their focus on supporting the analytical process by viewing the data along natural business dimensions and supporting iterative querying in which the result of one query can be the basis of the next. Additionally, the past several years have brought more widespread adoption of discovery technologies, such as data mining, in which the sophisticated algorithms can uncover patterns, correlations and relationships in the data that would take too long to find using manual or other automated techniques.

I have described for you different technologies for business intelligence. In today's architectures, we tend to pigeonhole users along these same categories. But there is something wrong with this picture. The technology categories should not drive the knowledge worker. Instead, the business problem should drive the knowledge worker. Instead of being a "retrieve" user, the knowledge worker should make use of whatever access technology or combination of technologies will provide the answer most efficiently.

This hypothesis seems obvious enough ­ so why isn't the world thinking along these lines? The constituent technologies have evolved independently. Discovery technologies have evolved from the machine learning, artificial intelligence and statistics communities. Retrieval technologies have grown from the database world. Now we have knowledge management technologies evolving independently. Vendors have contributed to this separation through their messages focused on their subset of categories.

The result is that the state of the technology today does not easily allow the use of business intelligence in an integrated fashion. Analysis tools and discovery tools have varying degrees of integration with the RDBMS in which our data is stored. Each tool may have its own preferred schema for storing the data. All of the tools have their own set of meta data to describe and catalog the data available for analysis.

The business intelligence vision of the future is a single meta data environment from which we can first choose the data for our analysis and then choose how to perform the analysis. This idea moves the focus from the analytical functionality to the content. The business intelligence environment will understand the storage of the data and deliver it to the analytical engine we need. As we fold knowledge management into this vision, the data available includes not just the fields from the warehouse, but documents, e-mail or even external data from the Web. Through the meta data, this unstructured data will be associated with the data warehouse.

What would this integrated business intelligence environment look like? Here's an example: A knowledge worker in the checking department of a bank who notices that there is unusually high deposit activity in the western region could drill into the western region and find that the activity is located around northern California. Considering that technology firms may be providing bonuses to their employees, the knowledge worker decides to build a model to predict investment purchases by the customers in northern California. The results of the model are displayed visually to the worker. The customers predicted as likely to purchase investments are identified by the model, retrieved from the warehouse and forwarded to the marketing group. In this scenario, the knowledge worker is not hindered by stovepipe analytical functionality. The worker makes use of reporting, OLAP, mining, visualization and access tools to fully discover and answer the business question.

We are taking small steps toward integrated business intelligence. XML may provide a framework for integrating the analytical technologies in the future, but this will take commitment to develop and share any standards. Portal technology offers the ability to integrate output from business intelligence functions at the user interface. While we are beginning to see a change in focus to the content, there is still much to be done to move the integration point back into the meta data and prior to the analytical functions.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access