Continuous Improvement - Data Quality

From MIKE2 Methodology

Contents 1 Activity: Continuous Improvement – Data Quality 1.1 Objective 1.2 Major Deliverables 1.3 Tasks 1.3.1 Conduct Ongoing Data Quality Monitoring 1.3.2 Associate Data Quality Issues with Root Causes 1.3.3 Execute Issue Prevention Process 1.4 Core Supporting Assets 1.5 Yellow Flags 1.6 Key Resource Requirements

Activity: Continuous Improvement – Data Quality

Objective

The Continuous Improvement of Data Quality covers the ongoing quantitative analysis of data assets. Data profiling and measurement software provides valuable insight into information quality issues in the existing information assets. However, they measure only the symptoms and do not provide significant insight into the root cause of poor quality data. Instead, the tool-based results need to be examined and root-cause analysis performed. Root cause analysis is the most creative phase of improving Data Governance and requires significant collaboration between the business and information technology to re-engineer processes, recommend improvements to architecture and execute steps for issue prevention.

Major Deliverables

• Improved Data Quality

Tasks

Conduct Ongoing Data Quality Monitoring

Objective:

Profiling results and other metric data gathering will be used as the basis of root cause analysis. Profiling should be conducted periodically – at least monthly – and the results are stored in the metadata repository and published into a dashboard report for tracking over time. Data Quality Monitoring can also be "operationalised" as part of the integration process to detect Data Quality issues each time they pass through the system.

Input:

• Operationalised Data Profiling Rules
• Data Governance Metrics

Output:

• Periodic Data Profiling Results

Associate Data Quality Issues with Root Causes

Objective:

Once issues are identified, representatives from the Data Governance team should convene to assess plausible root causes for the problems that have been identified. These issues a tracked within a tool, with a historical view associated with the root causes of problems, how problems relate and their impact on the business in term of cost/value/risk.

Input:

• Periodic Data Profiling results

Output:

• Identification of Root-Causes for Data Quality Issues

Execute Issue Prevention Process

Objective:

Identified issues are fed into the Issue Prevention process to assess the impact of correction and to determine whether the issue should be addressed or not. Issues should be addressed based on the cost/value/risk model that was defined earlier.

Input:

• Identification of Root-Causes for Data Quality Issues
• Data Governance Issue Tracking Process
• Data Governance Issue Prevention Process

Output:

• Recommended revisions and changes to those issues that lead to poor Data Quality

Core Supporting Assets

Yellow Flags

• Inability to associate data quality issues with root causes
• Failure to reduce root cause issues - data quality programme only involves fixing historical issues, not preventing them

Key Resource Requirements

• Information Architect
• Data Stewards
• Data Owners