An Alert Management Approach to Data Quality
By Joseph Bugajski,
Visa International and
Robert L. Grossman,
Open Data Group
www.opendatagroup.com
Abstract
Organizations have accumulated, automated and continue to generate and amass incredible amounts of transactional data. The amount of transactions occurring in the ether, and business dependency on data quality, continues to grow exponentially. However, traditional data mining models are limited in their ability to detect, validate and ameliorate problems identified in large sets of heterogeneous, distributed data.
Visa International Services Association (Visa) is an example of such an enterprise dependent on a robust data quality program. They utilize an alert management approach, an end-to-end framework that integrates strategy, governance and strong analytic models necessary to build and deploy successfully data quality program.
The alert management approach to data quality aligns the dimensional model of data quality with the business objectives and investigative process.
Architecture of an Alert Management Approach involves:
- Identification of business problem or opportunity
- Building statistical, rule-based, and architectural models
- Deployment of the models into production environments
- Investigation of alerts generated
- Continuous oversight and improvement of process
Learn how Visa's adoption of the approach has resulted in early, meaningful and actionable alerts over very large, distributed and heterogeneous data sets. Review the process for building an alert management approach, and examples of problems detected and resolved that deliver real business impact.
Download White Paper
|
