Imagine this: analytical research in the hands of research analysts
Only those who care count caribou
British Columbia covers an area of 94 million hectares (roughly three times the size of California), 90% of which is under state crown ownership. The provincial government’s caretaker of this land is the Ministry of Environment, Lands and Parks (www.env.gov.bc.ca). As part of its stewardship, the ministry is tasked with a surprising amount of data collection and analysis, ranging from counting permits for waste dumping and pesticide spraying to counting caribou and salmon.
"We’re not bottom-line value-driven; instead we focus on things like water quality," said Andrew Faulkner, technical architect of the ministry’s data warehouse.
However, the agency faces many of the same challenges as a profit-driven enterpriseÑdoing more with less. Reductions in payroll have meant reductions in staffing, leaving fewer employees to analyze data. Wherever possible, the ministry has been automating its business processes and has found business intelligence tools to help.
Analysis is run for employees at all levels of the agency. Requests are as diverse as park rangers looking at the number of people evicted from a park for rowdy behavior, or biologists seeking patterns in animal migration. Information collected throughout British Columbia is sent to an Oracle (www.oracle.com) data warehouse in Victoria where eight regional offices feed the system data on everything from radio collars on caribou to transmitters in fish.
With effective reporting tools, that information can be delivered in a summarized, usable format. The advantage is that analysis can be done by the people who ask the questions, as opposed to having to rely on IT.
"We’ve been able to move analysis to usersÑbiologists and park managersÑwho know about their business," Faulkner said. "These users just ask the questions."
Oracle’s ad hoc query and analysis tool, Discoverer, allows the ministry to access data warehouses, data marts and operational databases. It features query prediction and automatic summary capabilities.
"Business intelligence is often a set of questions that lead to the next question," said Vince Casarez, VP of Tools Product Marketing with Oracle. "Tools like Discoverer allow users to drill down."
Using Oracle’s integrated modeling environment Designer/2000, the ministry creates client-server applications using module components. Data input at any stage of development is stored in a central repository, enabling easy project management. Specific analysis can then be exported into Oracle reports where field staff access the data and bring it into a decision system.
"Our biggest problem is bandwidth moving data around," said Faulkner.
A problem with business intelligence isn’t a need for information but finding logical ways to summarize it.
"There’s a ton of data," Faulkner said. He cited environmental sampling data, which includes such things as pH and nitrates levels, and can contain as many as 2.5 million rows in a table.
"When developing BI systems, look at who users are," Faulkner advised. "Push the ability to quantify right down to the field level person. We’re empowering workers. We’re pushing decisions to specialists in the field instead of a bureaucratic infrastructure."