Integrating OLAP analysis and analytic components into day-to-day business applications has always been difficult, with most BI vendors offering stand-alone analytic tools. With the introduction of Crystal Analysis 9.0, Crystal Decisions now offers an OLAP reporting and analytic tool tightly integrated with its flagship product, Crystal Reports 9.0. Now developers can build multipage analytic reports and applications that feature guided analysis and navigation.

Under the Hood

From a user interface perspective, very little has changed between versions 8.5 and 9.0 of Crystal Analysis. The majority of enhancements deal with the way Crystal Analysis interacts with the underlying online analytic processing (OLAP) data source. Crystal Analysis 9.0 now lets you build analytic reports from data held on a number of OLAP servers, including Microsoft SQL Server Analysis Services, IBM DB2 OLAP Server, Hyperion Essbase, SAP BW, and Crystal Analysis Holos.

Using Crystal Analysis, you can create multipage documents that display OLAP data using a number of different elements, including an Excel-like worksheet with various charts or graphs and components to help you navigate your OLAP data. End users will spend the majority of time working with the worksheet, slicing and dicing OLAP data. The worksheet interface is easy to learn and users should have no problems finding the information they need.

For analyzing OLAP data, the worksheet provides all the tools you would expect, such as Top N, linear regression, and moving average. Users can also create their own custom calculations, but users unfamiliar with OLAP terminology will have trouble moving beyond simple calculations. The interface for editing calculations is also very rudimentary and could be better organized.

Several other features are hidden away within the spreadsheet: You can highlight values within the worksheet, making highlighting key information (such as when actuals don't meet budget figures) and exceptions easy to configure and understand. Another hidden gem is the inclusion of summary statistics that show the mean, standard deviation, and so forth, via a right-click menu without creating custom calculations. For complex financial modeling applications or users with a statistics background, these statistics are especially handy.

Crystal Analysis bundles the same engine found in Crystal Reports for displaying data in charts or graphs, and it provides the majority of standard chart types (such as pie, line, and graph) that let you drill down, drill up, and so on. But just like in Crystal Reports, the integration between the product and charting engine is a point of contention. If you've used Excel and want to duplicate the same charts and graphs in Crystal Analysis, you can create most of them but might have trouble controlling their finite aspects. (See Figure 1.)

Another graphing "mixed bag" is that all the objects (worksheets, charts, and so on) that appear on a page are tied together. If you have a page that shows both a worksheet and chart, changing the dimensions within the worksheet makes the chart also reflect that change. Although this trait is handy when you want to create a page around a single area (such as sales by salesperson), the trend in analytic apps is to show unrelated information on a single page or dashboard, which this feature doesn't permit.

In addition to worksheets and charts, many navigation components for analyzing OLAP data are available in Crystal Analysis 9.0. You can add either a Slice Navigator or Dimension Explorer to a page that gives you an alternative method of navigating OLAP data. When you want to display a single chart on a page, you can use these components to navigate your OLAP data without adding a worksheet.