Analyzing the OLIVE Branches
As the volume of data that companies collect continues to grow, managers struggle to understand how they can use this information to create a sustainable competitive advantage. One approach is to build a decision support application. What I see is users also beginning to demand a richer visual analytical environment. There is a new, emerging category of enterprise data visualization applications, which I term on-line visualization for the enterprise (OLIVE). OLIVE systems are chart-centric applications that deliver visual business intelligence to the enterprise. There are 12 attributes that an enterprise charting application tool should have to qualify as an OLIVE tool.
Dynamic, Interactive Charting
OLIVE systems deliver dynamic, interactive charting to the desktop. OLIVE users can manipulate the display by modifying fonts, changing chart type, or by adding color, labels, annotation, and titles. A rich set of charting objects enables the user to embellish a basic chart and add complex structures, such as multiple chart types, on the same display.
Chart Definition Language
An OLIVE system is based on a chart definition language (CDL) that can define both the content and structure of the chart. The content portion of the CDL defines the data or data sources for the chart, while the structure portion defines the chart attributes. A chart definition language file can be used by a chart rendering engine in the same way a Postscript file is interpreted by a Postscript engine to render a printed page.
An OLIVE application can accept data from virtually any source. This is a key difference between chart-centric software and data-centric applications, such as spreadsheets or OLAP tools.
An OLIVE system must be able to blend data from multiple, heterogeneous sources into a single, unified view. The data source responsible for each portion of the business graph should be transparent to the user.
An OLIVE system can accept live data feeds, allowing users to monitor real-time processes.
An OLIVE system can display a chart on different output devices -- browser, workstation, printer, pager, PDA -- using the appropriate rendering solution.
An enterprise-level system must be based on a server. Desktop-based charting tools lack the scalability, interoperability, security, or robustness required of an enterprisewide application.
Open Systems, Standards-Based
An OLIVE application must support industry standards and open platforms, such as Unix and NT, as well as data interoperability standards, such as ODBC and Java Database Connectivity.
An OLIVE developer needs tools and procedures for creating the charts, while the system administrator requires tools to facilitate the management of the charting applications once they are deployed. An OLIVE system provides support for the development, deployment, management and maintenance of chart-centric applications both within and between enterprises.
In a large enterprise, different users need to view the same data differently based on their profile or corporate role. For example, the CEO and CFO need to view and analyze corporate data at the enterprise level. Simultaneously, line-of-business managers must analyze similar data in their areas of responsibility. Consistency and data integrity constraints require that all use the same data set and the same chart definition, with the display limited by their profile attributes.
No data analysis software can be considered for enterprisewide deployment that does not support the major desktop personal productivity tools. This integration should be done using open desktop standards, such as OLE and ODBC.
Support Web and Windows Environments
An enterprise charting solution must leverage Internet-based technologies. This enables charts to be defined once and rendered everywhere. Managers can install the application on a Web-enabled server and manage it from a central location. Security control, data reliability, backup/recovery, performance measurement and capacity planning are all easier and cheaper in a server-based Web environment.
The chart rendering engine should support all the important Web media formats, such as HTML pages, Java applets, and GIF files, as well as other important formats that can be rendered by a browser applet or plug-in, such as BMP or PDF files.
If you want to learn more about enterprisewide data visualization, I suggest you check out the ChartWorks web site at www.chartworksinc.com. -- Robert Craig is director, Data Warehousing and Business Intelligence Division, at Hurwitz Group Inc. (Framingham, Mass.). Contact him at firstname.lastname@example.org or via the Web at www.hurwitz.com.