Data Mining Mortgage Documents for better Lending and Banking

It’s an old adage in the information management market that 80% of an organization’s data is unstructured. Historically, that has primarily meant documents. Today, just because organizations are applying analytics tools around their structured data, doesn’t mean their documents have gone away. Often, the data in those documents can be just as valuable as, and maybe even more important than, an organization’s structured data. Wouldn’t it be valuable if the lender could mine this document information? Say, for example, the lender wanted to chart historical housing pricing trends in different areas of the country to better predict future hot areas for mortgages? Or, what if they could better assess risk based on factors like annual earnings trends associated with loans that failed? Or what if they could mine their loan documents for unusually high instances of the same identification information, like a social security number, that might indicate fraud?