Web speeds safety data to Coast Guard

| SPECIAL TO GCNThe Coast Guard's Marine Safety Systems Development Branch has developed a Web decision support system to analyze maritime information and assist management in formulating effective safety programs.The Mission Analysis and Planning system uses an Oracle Corp. database as a repository and employs statistical analysis and data exploitation technology from SAS Institute Inc. of Cary, N.C.'MAP has made field users far more conscious of data quality problems,' said Lt. Cmdr. William Balsinger, the MAP project officer. 'As well as helping us to analyze information and adopt more efficient safety measures, we've seen a significant data quality improvement at the point of data entry.'Operating under the Transportation Department, the Coast Guard's primary goal is to ensure maritime safety. It investigates marine casualties and accidents, and conducts vessel inspections. The Coast Guard also responds to oil and pollutant spills at sea.With millions of vessels and vast seas to monitor, huge quantities of data are entered into the USCG system. Until recently, however, collected data was not easily accessible.'Field users lacked ready access to the information they were entering into the transactional system,' Balsinger said.Under the old system, users made requests to data analysts at Coast Guard headquarters, he said. Because there are a limited number of centrally based analysts, access to information often was delayed. Immediate access to safety information can mean the difference between life and death at sea, so the Guard partnered with SAS to redesign its data analysis procedures.MAP was designed to allow easy access to marine safety data by all levels of personnel. Initially, SAS engineers focused on replicating the existing system into a Web environment. That has opened up data assets to those who need them most, Balsinger said.MAP lets Coast Guard analysts rapidly identify safety and environmental risks in port areas. Once these risks are identified, local performance goals can be developed based on the risk factor. Over time, the system tracks progress in reducing or eliminating these potential dangers.The system consists of:OLAP software transforms data from a variety of sources into various dimensions, such as product line, geographical zone, market sector or time period. A user can synthesize any combination of these dimensions to either perform an in-depth analysis or create an ad hoc report on a specified area.OLAP is usually implemented in a multiuser client-server mode and offers rapid response to queries, regardless of database size or complexity. An OLAP server is a high-capacity, multiuser manipulation engine designed to support multidimensional data structures.Two servers are used for the MAP implementation. One functions as a Web server and houses the MAP Web interface, training programs and the online user manual, Balsinger said. The second server acts as an application server, housing the SAS software and the Oracle data warehouse.Multidimensional ad hoc reporting is an integral part of MAP. Users can view critical marine safety information against many different parameters, such as time frame, location and incident elements. For example, analysts can view oil spills by time, location, source and oil type. Once reports are generated, users can view the data in many ways, letting them conduct conditional analyses until they establish a pattern. As a result, commonalities can be spotted rapidly, letting the Coast Guard more tightly focus its safety efforts.'Local analysts can now view oil spills within their area of responsibility for the past several years and see how they compare to overall national trends,' Balsinger said. 'Alternatively, they can review personnel casualty data to determine where and why most of these casualties are occurring in order to focus our safety improvements in specific areas.'MAP has also assisted the Coast Guard during its annual review of its overall business plan, from which a five-year strategic safety plan was drafted. Two current goals, for example, are reducing passenger injuries and deaths on U.S. vessels, and reducing maritime oil spills.The system's capabilities include time series forecasting and regression analysis. Time series forecasting lets analysts review time-related data'such as the volume of oil spilled by month, quarter or year'and make predictions. Regression analysis complements this and makes it possible to identify measurable factors that could lead to a particular outcome.For example, a comparison of how spills correlate to billions of gallons shipped could help analysts predict how a 10 percent shipment increase would affect the volume of oil spilled.