News
 Developing
a Universal Approach to Cleaning Consumer and
Product Data
Wikipedia defines data as “When data becomes
organized, it becomes information.” It sounds so
simple, so why does data baffle us? The truth is
data quality is not a “one-size-fits-all” solution.
According to Colin White, Founder of BI Research,
data quality needs to be an ongoing
project-by-project basis.
In a Business Intelligence Network web seminar,
Colin White spoke of the many contrasts of data, and
how every type of data has a solution.
Types of Data
• Metadata Definitions & Rules
Technical & business information about data and its
enforcement)
• Master Business Entity Data
Key business entity data (products, customers, org.
structure, chart of accounts)
• Master Reference Data
Key reference data (code & lookup tables,
taxonomies)
• Activity Data
Business data created by transactional, BI, and
collaborative applications
Ensuring Consistency: Data Profiling and
Cleansing
• All data must be syntactically consistent with its metadata definitions
and rules (data type, data codes, data format)
• Activity data should be unambiguous and semantically consistent with its
business definition (records and fields contain the
correct business data)
• Activity data must be clearly identified and consistent with its
associated master data
• Data profiling tools can be used to assess the quality of existing data
• Data cleansing tools can be used to validate and improve data
consistency –can be used proactively or reactively
Data Accuracy Considerations
• This is the most difficult aspect of data quality governance and
assurance
• Improving data consistency enhances data accuracy
• Data accuracy is also highly dependent on the users and applications
that create and use it
• Auditing tools and data lineage information help in assessing and
validating data accuracy
Take Aways
• Achieving data quality is not an all or nothing project—if you think it
is then you will fail
• Amount of data and content is increasing rapidly—a high percentage of
this data is not fully assured, but it is still
useful
• Different users need different levels of data quality—categorize data by
business need and required confidence levels
• Managing master data and processing unstructured content are important
trends
• Better to have an integrated data cleansing technology architecture than
deploy piecemeal solutions
• Data quality management is an ongoing process that requires the constant
monitoring of all types of business data
---Sources:
Colin White, Founder of BI research and Business
Intelligence Network (www.B-eye-network.com).
|
|
|
Melissa Data
|
 |
| Enhance your
website, software or database with
easy-to-integrate data quality programming tools
and web services. |
|
|
|
|
 |
|
Save money on postage using leading
mail preparation software and other
direct marketing products. |
|
|
|
|
 |
Update & standardize addresses and
find out more about contacts in your
database.
|
|
|
|
|
 |
Find new customers perfect for your
business with our online and
specialty mailing lists.
|
|
|
|
|
 |
Locate the business information you
need such as ZIP Codes, address
verification, maps.
|
|
|
|
|
