Data Quality - New Methods for Improvement!

Data Quality is a measure of data set accuracy. Data sets lacking in quality are less valuable than data sets being of a high quality. As has been stated many times, data inaccuracy can be very costly to an organization. In addition, data sets usually become less accurate over time. The original data becomes stale! This is understandable as in the real-world people relocate, countries split, companies merge, and sometimes companies go out of business. The data sets must be able to reflect the changes both chronologically and in the recasting of time.

In order to maintain the value of a data set, organizations often institute a process of data cleansing. This is done especially for integrated decision support systems were extra value may be drawn from data collected to support operational systems.

Tandem Primary Keys ^tm was originally developed to support the data quality improvements that impact multiple data items in a single data set. These are improvements, such as removing duplicate data items, and merging multiple data items that now represent a single data item. Prior to Tandem Primary Keys, data often needed to be extracted and transformed in an attempt to restore the accuracy to such data items. These transformations are very expensive to develop and only allow a one-shot attempt to improve data quality. The resultant data only provides the current recast view of the data while the chronologically accurate view is lost!

Tandem Primary Keys supports the cleansing of data without extraction and transformation. Instead of a one-shot ETL process, a data set may now be cleansed within its original database anytime a data steward deems a data quality improvement is needed. In addition, you do not lose the chronologically accurate view of the data set or the ability to recast time in various ways. These data changes are also completely reversible while data transformations normally are not.