In-Depth

SPSS 14.0 Targets the Enterprise

SPSS introduced a new version of its flagship data mining and statistical analysis product the company says is its most enterprise-ready tool to date

• By Stephen Swoyer
• 08/24/2005

In certain circles, two companies have long been synonymous with what is today called business intelligence (BI): SAS Institute Inc. and SPSS Inc. Both companies have been shopping data mining and statistical analysis tools for the better part of 30 years, and both boast exceedingly loyal customer bases.

Not surprisingly, both also have aspirations beyond their bread-and-butter technology spaces: SAS launched its much-ballyhooed SAS9 BI “suite” 18 months ago, serving notice to upstart competitors such as Business Objects SA, Cognos Inc., and Hyperion Solutions Corp. that it had arrived.

SPSS, too, has expanded its own focus beyond the insular data mining and statistical analysis markets, first with its popular predictive marketing solution for business users (“Clementine”) and—over time—building out the capabilities of its popular statistical analysis suite.

Earlier this month, SPSS introduced a new version of this product, SPSS 14.0, touting a host of improvements that the company says makes the program its most enterprise-ready tool to date. “What we’re looking at here is we’re broadening the types of analyses that people can do with SPSS,” says Nancy Dobrozdravic, vice president of product marketing with SPSS. “For example, you can take any kind of third-party scripting language like a Python or Perl. You could have that as part of a larger job, you may even call in things from another software package, that can all be invoked within SPSS.”

At the same time, Dobrozdravic says, SPSS has opened up its output—in this case, in the form of programmable objects. “This means I can take the results of an analysis and take that as input into another SPSS job or into another application for general processing. You can use us as home base, if you want, but this helps [SPSS 14.0] fit better into the overall working environment.”

As a result, Dobrozdravic claims, SPSS 14.0 is no longer the highly insular platform of old. For example, she points out, the revamped SPSS also ships with a common set of services—including unified security management and a common repository—used by all SPSS products, such as the company’s Clementine data mining tool. “We have been working for the past couple of years in putting together a very robust services-oriented architecture that provides services such as security, scheduling, notification, authorization, and probably premier [among] those is a common repository.”

All of these changes have the effect of opening up—i.e., exposing—SPSS to other applications and expanding its user constituency. Support for scripting is only the beginning, however, says Dobrozdravic. “Down the line, we’re going to put in APIs [to facilitate access to SPSS] so [customers] can drive SPSS with custom user interfaces. In the larger enterprises, where you really get the business user and the business consumer, this is part of the whole idea of making SPSS part of this larger services infrastructure.”

Improved Data Mining, More Flexible Licensing

SPSS 14.0 includes the SPSS Base 14.0 product and associated add-on modules, along with SPSS Server 14.0. New in the SPSS 14.0 release are several additional data mining algorithms, support for more file types—including OLE DB data access and SPSS’ own Dimensions products—improved data validation and anomaly detection capabilities, a new Chart Builder user interface (that lets inexperienced users build custom charts), and support for SPSS’ Graphics Production Language (which advanced users can tap to build custom charts). SPSS 14.0 also addresses a long-time shortcoming in the SPSS universe—namely, the inability to work with multiple data sets in a single SPSS session. As of SPSS 14.0, this limitation is, effectively, history.

There are other goodies, too—especially in the area of licensing flexibility. SPSS now supports a network commuter license (which Dobrozdravic says is designed for off-site or “on-the-road access”) that lets users work even if they don’t have a persistent network connection. Other new licensing features include server-side support for Unix and Linux, along with Windows.

What it all adds up to, says Mike Schiff, a senior analyst with consultancy Current Analysis Inc., is an important new release that further fleshes out SPSS’ enterprise story. “By integrating its products and sharing a common set of services with other SPSS products … SPSS 14.0 should raise user productivity and efficiency—and thus, the overall analytical effectiveness of the organization,” he writes. In part because of its popularity as a desktop tool for statistical analysis, SPSS hasn’t always been taken seriously as an enterprise solution. That should change with the new release, says Schiff.

“SPSS 14.0 allows the company to position itself as an enterprise-class offering capable of meeting the needs of the enterprise, while also reinforcing its already strong desktop position,” he writes, adding that SPSS 14.0—which shares a common set of services with SPSS’ Clementine predictive analytics server—should reinforce “SPSS’ predictive analytics initiative while demonstrating the continuing evolution of the company’s technology foundation.”

In this respect, Schiff lauds the company’s new open API. “By exposing its backend API, the company opens up the SPSS engine to third-party programming languages and greatly enhances its ability to be integrated with third-party applications. This openness is further supported by the existing ability to produce XML output,” he says.

Given the emerging importance of performance management and forecasting, Schiff argues, SPSS should consider touting other new version 14.0 features. “SPSS should highlight the bulk forecasting capabilities of the new Expert Modeler feature in SPSS Trends 14.0 and stress that it, too, has ‘high-performance’ forecasting technology.”