Clean Data is a Key Component of Automated Value Estimates.(Brief Article)

Case Shiller Weiss is not the only firm that claims to have the largest, or at least one of the largest, property record databases in the country. But Terry Loebs, a vice president with the firm, says size is not the only thing about the database that matters.

"If nothing else, I'm convinced that we have the cleanest," he said.

As the old saying goes, garbage in; garbage out. For an automated property valuation model to be useful, the data inputs have to be accurate, he said. But CSW's clients make major portfolio investment, sale and purchase decisions based in part on the home value data, so the company is willing to put a little extra effort into data scrubbing.

"We take very seriously the non-glamorous side of this business," Mr. Loebs said.

CSW (www.cswonline.com) relies on multiple sources of public records regarding property sales. About 20 companies that compile assessor /recorder information from public records feed data to CSW, which uses repeat sales information to build its Case-Shiller Home Price Index and related products.

Mr. Loebs acknowledges it would be easier to work with one national data source and have only one contract and system to administer, but CSW wants diversity and redundancy in order to catch errors and clean up the data. By "cross weaving and overlaying" information from multiple sources, CSW is able to identify and weed out suspect data.

CSW uses automated filters to reconcile the raw data over time and across data sources.