Data Grading Project
data quality effort
A leading reinsurance company utilizes sophisticated catastrophe management software to analyze the property risk represented by client insurance portfolios. What this third party software doesn't do is to grade the quality of the data it processes. This left our client exposed to losses stemming from portfolios whose risk could not be properly assessed.
Again we designed a domain specific language, using spring.net, Visual Basic and MS SQL Server, for grading the quality of insurance portfolio datasets. Our application scores datasets by looking for suspicious records: incompletely specified ones and ones containing default or nonsensical values. The scoring rules are encoded in an xml based domain specific language allowing our client's employees to transform, extend, and create analysis without the need for additional programming.
