SAS Institute is set to announce a new project aimed at helping companies link the various sources of operational data.

At its SAS Users Group International (SUGI) conference in San Francisco, SAS will roll out a new customer data integration server and detail new features in its enterprise data integration server, said Jim Goodnight, the company's CEO. SAS will expand the data-quality features in the enterprise data integration server for profiling, monitoring and cleaning data, he said.

"It's a shame to wait until people want to use data in a warehouse before cleaning up the data," Goodnight said. "It should be cleaned up in the operational system."

SAS said that the new customer data integration tools will be designed to allow companies to synchronise, consolidate and manage customer information from across an enterprise.

Debbie Weisensee, manager for a BI project at McMaster University in Canada, said her organisation has been using the SAS data-integration server as part of an ongoing effort to build an enterprisewide data warehouse. The university did an assessment of data quality at the beginning of the project and changed some data definitions to ensure that they were consistent across its campus. However, some records come out of transactional systems with missing fields or other problems and have to be rejected before being added to a data mart, she said. Now the university has to do some extra programming to check for certain data-quality requirements. Weisensee said she is interested in learning more about the new features in the integration server to possibly eliminate some of that programming.

In addition, the university in September will be rolling out SAS performance management software to create scorecards, another project where the quality of the data in the warehouse will be critical, she said.

Keith Gile, an analyst at Forrester Research, said the data integration initiative shows that SAS is hoping to expand the definition of what is included under the BI umbrella to include support for the data warehouse.

"This is extending BI to be an information platform ... not just the reporting and analysis part of BI," he said. "It shouldn't be, 'Let's get it out to the BI tools before we realise it's dirty data.'"