Blog: data quality

Showing blogs: 1–3 of 3

The Necessity of Data Profiling

Sep 21 2009

Data profiling is the process of analyzing data sets to gather information on data types, statistics, patterns, relationships and other attributes.  This process is a critical task often scheduled just prior to the start of a design phase for a new database initiative.  Typically, data profiling is seen primarily as an input for data modeling and ETL design to identify data types and entity relationships.  The value of thorough data profiling actually extends far beyond these tasks and may help steer the direction of the project and ensure its success.

Below are some of the ways data profiling can add value during a new database initiative: 

Continue Reading

Data Quality - Three Essential Links

Jun 24 2009

  1. Magic Quadrant for Data Quality Tools (Gartner) -- A good starting point if you want to become familiar with the main vendors in the data quality space.
  2. Surprising Value of Data Profiling (Kimball - PDF) -- Ralph Kimball provides a concise explanation about the value of data profiling, one of the essential components of a data quality initiative.
  3. 7 Sources of Poor Data Quality (Melissa DATA) -- William Mcnight identifies seven sources of data quality issues:
    • Entry quality
    • Process quality
    • Identification quality
    • Integration quality
    • Usage quality
    • Aging quality
    • Organizational quality

Continue Reading

Managing Customers’ Data Quality Expectations

May 13 2009

A recent client concern was raised when an unusually high number of mail pieces were returned as undeliverable by the US Post Office (USPO).  The negative impact to the customer is obvious: Not only did they spend good money on postage to mail the piece in the first place, but now they have to expend internal resources to determine the correct addresses, resubmit the documents to the USPS and field customer concerns of late and/or missing mail pieces.

Having just run the NCOA (National Change of Address) updates for the client, it was natural that the client wanted an explanation why the NCOA processing did not catch more of the undeliverable addresses.

Continue Reading