Page 104 - DCAP606_BUSINESS_INTELLIGENCE

Unit 6: Business Intelligence Project

Notes
Standard Deviation
Variance
Consistency within a database is another important factor to determine through data
profiling. For example, there may be an order table which should only have orders for
customers established in the customer table. Perform queries to determine whether this is
true.
Improve Data Quality
Data profiling may reveal problems in data quality. For example, it might show invalid
values are be entered for a particular column, such as entering ‘Z’ for gender when ‘F’ and
‘M’ are the valid values. Some steps that could be taken to improve data quality include:

Work with data owners to define the appropriate level of data quality. Build this
into a data governance program.
Determine why there are data quality problems — do a root cause analysis.

Correct the data in the source system through manual or automated efforts.
Add edits or database rules to prevent the problem.
Change business processes to enter correct data.
Make data quality visible to the business through scorecards, dashboards and reports.
Save Results for Further Reuse
The information gathered during the data source discovery process is valuable metadata
that can be useful for future data warehousing or other projects. Be sure to save the results
and make available for future efforts. This work can be a great step toward building an
improved data resource.
Question:

Discuss the case study in contrast with efficient and effective workflow of obtaining the
right source data and using it in the data warehousing and business intelligence project.
Source: http://infogoal.com/datawarehousing/data_sources_2.htm

6.6 Summary

A multidimensional model must contain at least one data source object, but you can add
more to combine data from several data warehouses.
A data source connection can occasionally use Windows authentication or an authentication
service provided by the database administration scheme, such as SQL Server authentication
when connecting to SQL Azure databases.
The attachment string is formulated based on the properties you choose in the Data Source
Designer or the New Data Source Wizard.
After you have defined the data sources that you will use in an Analysis Services project,
the next step is generally to define a data source view for the project.
The ability to use the DataView to modify data in the underlying table is controlled by
setting one of three Boolean properties of the DataView: AllowNew, AllowEdit, and
AllowDelete.

LOVELY PROFESSIONAL UNIVERSITY 99