1. Quality From The Eye Of Business Intelligence
“Quality at the heart of BI”
By: Kamel Badawy
2. Agenda
Introduction
What Does Quality Means to BI?
BI Data Quality Dimensions
Current Standing &Where to?
BI RoadmapTowards Data Quality
3. Introduction
Organizations are discovering that data quality deficiencies have
a significant impact on their most strategic business initiatives,
often holding them back from achieving:
Growth
Agility
Competitiveness
Transparency
In addition to challenges with growth and agility, compliance
and transparency pressures increasingly bring DATAQUALITY
issues to the fore — it is no longer acceptable to ignore flaws in
data, and organizations must prove the accuracy of information
that they report internally to top management or to auditors,
regulators and the public
BI Objective:
Is to Build clean, accurate & reliable data warehouse
BI should deliver data that is necessary for decision-makers
4. What Does Quality Means to BI?
Quality is critical to DataWarehouse and
Business Intelligence. Better informed, more
reliable decisions come from using the right data
quality technology during the process of loading a
data warehouse. It is important the data is
accurate, complete, and consistent across data
sources.
Data Quality is a multi-dimensional measurement
of the adequacy of a particular datum or data
sets. In business, data quality is measured to
determine whether or not data can be used as a
basis for reliable Business Intelligence and for
making organizational decisions.
REDUCTION
Cut BI project
failure rates in half
COST
Lower overall cost
of BI/Data
Warehouse
solutions
VISIBILITY
Implementing a
culture of
measurement
provides clear
visibility for all
parties
Why BI Data Quality?
5. Numbers & Facts
Of organizations
believe they’re
negatively affected
by inaccurate data
Of businesses admit
their data is not
accurate
35.00%
30.00%
30.00%
29.00%
23.00%
20.00%
11.00%
3.00%
0% 5% 10% 15% 20% 25% 30% 35%
LOW QUALITY, ACCURACY OF DATA
LIMITED DIRECT BENEFIT TO MY ROLE
DIFFICULTIES IN ASSESSING WHICH DATA IS TRULY
USEFUL
LACK OF NECESSARY SKILLS
PROBLEMS TO COMMUNICATE DATA
LACK SUFFICIENT EXPERTISE
ABILITY TO TAKE ACTIONS BASED ON DATA
PRESENTATION OF DATA IS IN AN UNUSABLE FORMAT
Barriers to integrating more data in decision making
25% of Critical
Data in the
World’sTop
Companies is
Flawed
How confident
organization are in
their data
6. BI Data Quality Dimensions
Data Quality
Dimensions
Description
Validity
Data accurately represents reality or a verifiable
source
Completeness
Records are not missing fields and datasets are not
missing instances
Integrity
The appropriate links and relationships exist among
data
Consistency
Data that exists in multiple locations is similarly
represented and/or structured
Uniqueness Data that exists in multiple places has the same value
Timeliness
Data is updated with sufficient frequency to meet
business requirements
Accessibility
Data is easily retrieved and/or integrated into
business processes
Data Quality
Dimensions
Description
Existence
Data reflective of meaningful events, objects and
ideas to the business has been collected
Usability
Stakeholders understand and are able to leverage this
data
Clarity
Data has a unique meaning and can be easily
comprehended
Believability Data is deemed credible by those using it
Objectivity
Data is unbiased and impartial and not dependent on
the judgment, interpretation or evaluation of
individuals
Relevancy
The data is applicable to one or more business
process or decision
14. BI Roadmap Towards Data Quality
Focus on the Right Things
Establish a clear line of sight between the
KPI/KRI impact of data and data quality
improvement
Use data profiling early and often
Design and implement data quality
dashboards for critical information such
as master data.
Fit for Purpose
Clearly define what is meant by "good
enough" data quality
Establish a data & report standards
across the organization
Move from a truth-based semantic
model to a trust-based semantic
model.