Panopticon Explorer .Net

Originally published online in a different form

Many tools aim to help business users display complex numeric data and aid in decision-making processes, such as business intelligence software from Actuate, Business Objects, Cognos and spreadsheet applications such as Microsoft Excel. Now Panopticon Software's Explorer .Net makes its mark with a visualization program that can be a powerful ally to analysts working with large numeric data sets, particularly those with structures easily mapped to correspond to components--subcategory, size and color--in the treemaps it creates (see "Treemaps Explained," page 30).Panopticon's software is available in three versions--Learning, Professional and Premium. The latter two can connect to enterprise resources, such as relational databases, and the Premium edition supports real-time data. Such features increase the tool's value in a number of fields in which large data sets must be understood rapidly. That includes IT, to study the use of system resources; engineering disciplines, to analyze the behavior of any number of complex systems; and marketing, to explore sales data across multiple regions or the impact of advertising in a state down to a particular zip code. There, the software's ability to jump between various levels of detail is an asset.

Treemaps, developed as a research tool at the University of Maryland, underpin some other analytical tools, such as The Hive Group's Honeycomb, which transforms data from a database into an information map viewable via a Web browser. The treemap Panopticon creates appears at first as a tangle of rectangles, representing a large amount of data densely packed into a small area. After a few minutes, though, you can begin to make sense of it. The ability to handle complex data comes at a price: You must spend time configuring the tool to your data set, and query results don't appear as a neat graphic to use in a PowerPoint presentation. The tool's abilities, however, excel in the hands of a competent analyst.

Growing a Tree

I created an Excel spreadsheet containing a fictitious investment portfolio. Each row contained information on one holding, including its current price. I opened the spreadsheet in Panopticon, but a treemap based on my data didn't immediately appear. I had to map the various fields in the spreadsheet to the different components of a treemap's structure--risk fields to colors, for instance. The application also was sensitive to the order of the fields in the original spreadsheet. If you want to establish category and subcategory relationships, the columns must appear in the same order as the hierarchy in the data.

After identifying and correcting for these sensitivities, I mapped the various fields from the spreadsheet into the appropriate treemap components. My category "Bond" included the subcategories "Municipal Bond" and "Corporate Bond." I mapped risks to colors, with high-risk investments displayed in red and safer investments in green. I mapped quantities and current values onto the size component, so that holdings in which I had a significant investment appeared as relatively larger rectangles.

With the mappings complete, I made queries such as, what high-risk investments represent a large percentage of my portfolio? With the data displayed I could see large red-filled rectangles representing the high-risk investments in my portfolio.

I was struck by the product's ability to show various levels of detail. I could have looked at all the data across all equities or drill down to individual stocks. Excel is cumbersome for analyzing data with category/subcategory relationships, because you cannot easily jump between various levels of detail. Once it's clear what data is required to answer a specific question using Panopticon, however, Excel provides a better platform for driving home the point using familiar charts and graphics.

Edward Hand is an independent software consultant in Madison, Wis. He has more than 15 years' experience as an IT analyst, developer and project manager, and has worked in defense, finance and other industries. Write to him at [email protected].

TreeMaps ExplainedTreemaps display data in three complementary ways: First, they show structure. A treemap effectively captures the categories and subcategories of a hierarchical data set and lets the viewer determine the level of detail appropriate for the task. Second, treemaps show size or area. Each subcategory or node is sized to a particular dimension; larger values get larger rectangles. Third, each rectangle is colored according to the value of another field within the data. These three techniques complement each other in that there is little interference between them; for instance, it's easy to pick out the placement or size of an element without its color confusing the meaning.