The Push for Better Database Tools
Though the early transactional and operational systems produced massive quantities of raw data, many decision-makers soon felt like prisoners of the systems because early attempts at tapping into the data stores in meaningful ways led to the discovery that the systems were not well-suited for providing decision-making support (McFadden, Hoffer & Prescott, 1999). This has spurred the demand for tools that transform raw data into meaningful information. Lack of integration between operational systems that contained relatively little historical data also presented a significant obstacle when attempting to merge and extract related data in meaningful ways, such as to analyze what if scenarios, track operational trends and support strategic plans. In addition, since operational systems focused on update-oriented transaction processing, their primary design goal was to optimize performance. Performance optimization was irrelevant in informational support systems, however, because they are nonvolatile and informative in nature making ease of access and usability more important design objectives requiring a different architecture (Inmon, 1992).
Recognizing the technological differences between operational and informational systems has resulted in the emergence of data warehouses designed to provide an integrated, company-wide view of high-quality information, which has further spurred demand for online analytical processing (OLAP), data-mining and data-visualization tools. Although OLAP systems have the ability to answer who and what questions in the same way traditional query tools do, it is the OLAP system’s ability to answer what if and why questions that sets them apart by enabling decision-making about future actions. OLAP transforms data into strategic information that can include calculations and analyses, such as time series and complex modeling, which enables decision-makers to move from data access to information to intelligence using interfaces that allow for basic navigation and browsing.
Where OLAP is ideal for slicing and dicing data to answer forward-looking what if and why questions, data-mining enables knowledge discovery through pattern recognition algorithms intended to explain past events, prove or disprove hypotheses, or explore potentially lucrative new data relationships that may be indicative of unexpected trends (McFadden, et al., 1999). According to Waldrop (2001), data-mining is expected to be one of the top ten technologies to change the world, turning what Fayyad (as cited by Waldrop) refers to as data tombs into true data goldmines. Ulfelder (2000) also observed, while data-mining is increasingly becoming the ideal backbone for customer relationship applications, data visualization tools are becoming standard features of statistical analysis software enabling users to move quickly from data analysis to knowledge synthesis. Data visualization software's ability to accentuate key data relationships helps combat information overload and makes significant and potentially threatening data conditions virtually impossible to miss and easy to understand.
Inmon (1992a) has dubbed the last decade of the 20th century the age of the data warehouse as analysts and decision-makers realize the enormous advantages of managing their organizations at the executive level. The first decade of the 21st century may prove to be the age of interfaces for data warehouses. However, many companies are shying away from customized interfaces in favor of standard ones that allow easy access to data by people who are geographically dispersed. Since the Web is rapidly becoming the preferred global computing tool, it provides the ideal interface for organizations seeking a means to work collaboratively from remote locations.
Currently, eXtensible Markup Language (XML) holds great promise to meet this need (Suciu, 2001). XML has gained wide acceptance as a standard not only for describing documents, but also as a data description language for all types of information. XML will eliminate the need for custom-built interfaces to access data; and with browser companies already supporting XML, browsers are good candidates for becoming universal user front ends. Work still needs to be done, however, to standardize how XML taps into legacy databases. Since, this is a rapidly advancing area, it is prudent to watch the developments closely.