Advanced Data Visualization: WebFOCUS Visual Discovery

By Rado Kotorov

If you have been following Business Intelligence trends, reading BI publications or attending industry events, you most certainly have heard about advanced data visualization, visual analytics or visual dashboards. All three terms are references to the visual display of large sets of multidimensional data in an easy-to-comprehend and manipulate manner.

As with any emerging technology, there is buzz and confusion about what it does, what differentiates it from existing tools, and, most importantly, what are the business benefits of adopting it. In this case, the reference term "advanced" does the technology injustice. "Advanced" implies that the main user will be the technically savvy business analyst – that 2 percent of BI users at the top of the user pyramid. But this is not the case. Advanced data visualization technology will appeal to the same users who like the WebFOCUS parameterized reports for their ease of use, and who are more or less intimidated by OLAP and multidimensional slice and dice.

Analogously to structured reporting, in which a single reporting template can be used to generate hundreds of reports, Visual Discovery allows users to develop a single template in which numerous dimensions and measures can be displayed and compared. From this perspective, Visual Discovery reduces the complexity of OLAP and, thus, empowers the less technical report consumer to perform more in-depth analysis. Let me explain why.

In the OLAP environment, analysis is driven by the selection of dimensions and measures from dropdown boxes. If you have three dimensions, each comprising 10 values, there are 1,000 measurable combinations. This model is actually very common — sales by year, region and product will produce this level of complexity. Yet, a robust business analysis would require the manipulation of at least seven to nine dimensions. No wonder many users feel intimidated by the intricacy of OLAP.

Furthermore, to make all possible dimensional comparisons the user has to "flatten" the OLAP (i.e., expand all dimensions) which will create a report with 1,000 measurable cells. Naturally, many users create smaller reports, and, then, manipulate the dimensional combinations to display the measures of interest within the smaller contextual framework.

However, the transition from one dimensional combination to another in a tabular report is very memory-intensive since users try to remember the values from the prior context and compare them against the values in the current context. This creates a psychological barrier for the non-technical user to adopt OLAP. Advanced data visualization has emerged as a means to resolve this issue and reduce the complexity of multidimensional analysis.

Visual Discovery, for example, reduces the multidimensional complexity inherent in OLAP by providing a common framework, in which all dimensions are displayed simultaneously on a series of charts linked to one another. Thus, instead of having a report with 1,000 measurable cells, the user from the earlier example is presented with three linked bar charts displaying yearly, regional and product model sales. Those three linked charts provide the common framework for sales analysis.

Conceptually, the difference with OLAP is that in Visual Discovery the user is presented the entire picture, while in OLAP the user sees fragments (slices) of the picture. Similarly to WebFOCUS structured reporting, the user never leaves the framework to perform a different analysis.

The linking of charts allows the user to see a different perspective and focus on different sets of

details without losing sight of the entire picture, much like rotating a three-dimensional object to examine all of its sides. For example, the selection of any particular year on the annual sales chart would change the other two charts from comparing total regional and product sales to regional and product sales comparisons within the year of interest.

Alternatively, the selection of any particular product would change the other two charts to annual and regional sales comparisons for the selected product. Since the user never leaves the framework, there is no loss of context, which in turn facilitates the detection of changes in the graph patterns and allows the user to quickly identify valuable comparisons and business trends.

So, when should you consider developing BI analytic visual dashboards?

In general, OLAP packs 300 percent to 400 percent more information than any report, and advanced visualization packs 300 percent to 400 percent more information than a regular dashboard.

Based on this, there are three key indicators that advanced data visualization may be beneficial to your users. First, your model has more than three dimensions that users actively navigate to slice and dice the data. Second, your user dashboards have more than two drillable charts or gauges, and users frequently drill on them to obtain different views of the data. Third, your users frequently drill down to the data details to discover exceptions, anomalies or other facts of interest.

Since there are many business benefits of visualizing the data details, we will discuss some of them in more detail in a follow-up article.