Sidebars: Putting BI To Work
Tracking Down a Sales Slump
Cognos' Anil Dilwari offers the example of finding a reason for a sales slump to explain how the different pieces of Cognos software work together to build a picture and track specific information.
In the Cognos scorecard, green means everything is OK, while red means there is a problem. Suppose you come in one morning, look at your data scorecard, and you see U.S. sales marked in red. You want to do further analysis and find out what went wrong, so you go into the Dashboard setup to see a graphical representation of the U.S. sales figures presented on a map and you see that California is shaded in red, indicating that it is a state with a problem. Clicking on California launches a multidimensional OLAP analysis environment. You find out that the main problem is in Los Angeles and San Francisco, and specifically it's a problem that surfaced in the last 30 days. You drill through to the lowest level of detail to a report outlining the status of the sales people in those two problem areas and you find that your two best representatives from Los Angeles and San Francisco have actually left the company within the last 30 days, which explains the reason for slumping sales in these areas.
You can take this a step further and get proactive by enabling event detection, so that if you ever lose two key sales people in this manner ever again, you will be informed by cell phone, pager, RIM device, or however you wish to be notified.
UC Berkeley Provides Unique Data Access
Debra Kelly, museum information specialist at The University of California at Berkeley uses Business Objects to put powerful query tools in the hands of end-users in the campus' Museum Informatics Project (MIP). Kelly says she bought Business Objects way back in 1995 after evaluating 30 tools that were around at the time. She was attracted to Business Objects because it put query writing in the hands of her non-technical users without them having to know any SQL (structured query language). She says users could simply drag and drop items they wanted to report on into the query panel. This appealed not only to her, but also very much to her end-user community.
Kelly's department is responsible for maintaining the MIP databases. She says that her two main Business Objects users are the Botanical Gardens and the History of Art department. The Botanical Gardens, which maintains a collection of over 35,000 plants, uses Business Objects software to keep track of items in their collection. For example, when they receive a request from students or researchers, such as all plants in a certain genus, they can build a query in Business Objects without help from IT and generate a printed report to give to the researcher. The History of Art department uses Business Objects to keep track of its collection, which includes more than a half-million slides. Although Kelly has developed some custom reports for her users, she says that for the most part, users can generate the reports they need without her help and that's why she continues to use Business Objects after all these years.
Pharmaceutical Company Mines Unstructured Data
Insightful's Coombs says their InFact product works quite well in the pharmaceuticals market because of how it integrates into the nature of their research process.
Coombs says a pharmaceutical company may do all kinds of research to see how a certain protein may impact a particular gene. They may do experiments and generate large amounts of data, leading to the development of a drug. They store this information in conventional databases, but once they find something promising, they need to mine the research literature to see if what they are developing has been touched before. This involves huge volumes of text in the pharmaceutical journals. It is nearly impossible to review this volume of data by hand, Coombs says, so they need an automated solution.
InFact ingests the text-based data and analyzes it looking for the user-defined text relationships to see if another company has made a similar discovery. In this case, it might ingest a huge database called Medline, which contains pharmaceutical and biological publications. InFact then performs information extraction, which means it reads all of the documents sentence by sentence looking for any information that shows a relationship between the gene and the protein this company was researching. At the end of the process, it generates a report of what it has found.
Companies Featured in this Article
Business Objects www.businessobjects.com
IBM www.ibm.com IDC www.idc.com
Shore Communications, Inc. www.shore.com