Metadata Management and
Data Quality

A Comprehensive, Intelligent Approach
For Ensuring Ongoing Data Quality

Setting the Standard

To improve customer relationships, predict better outcomes, and maintain compliance with a growing collection of legislation, companies are becoming increasingly focused on data quality.

When measuring “quality,” however, one must evaluate the adherence to, or deviation from, some defined standard.  What is the defined standard when measuring the quality of the data?  The answer is simple, straightforward, and often overlooked: metadata.

The Overvlooked Key to Data Quality

Metadata, in short, is data about data.  For many, metadata is simply the description of an element’s data class (e.g. string, integer, etc.), its precision, and/or its narrative definition.  But properly collected metadata should offer a much more comprehensive portrait of each structure, element, and value domain.

The Lightwell metadata model clearly and fully communicates data’s origins, logical and physical content, domain value decodes, population patterns, and, very importantly, applicability conditions.  By comprehensively documenting what the data should contain, Lightwell metadata is employed to evaluate the quality of what it contains.

Metadata is Not Just about Content

Most metadata efforts limit their review to describing what is in each table or field (e.g. data types, field/table names, field/table definitions, etc.). When it comes to content, however, properly collected metadata should not only focus on what but also when.

The Lightwell metadata approach ascertains when the content of each element or structure is applicable.

  • Element-Level Applicability

    Often, one can predict the circumstances under which a data element will be populated.  When preparing metadata, the Lightwell methodology ascertains when each attribute should and should not be populated.

    Applicability relationships are critical in distinguishing content that is missing because it is not applicable, from missing content that is unknown.  Applicability relationships are also important in safeguarding end-users from misleading information.  In a world of imperfect data, analysts may unwittingly encounter data elements that are populated, but not applicable.  Downstream analyses can suffer from misleading results.

  • Structure-Level Applicability

    Because relational data analysis involves joining structures, the Lightwell metadata methodology also ascertains the applicability of structural relationships.  Just as the preceding processes evaluated when an element should and should not present content, Lightwell’s metadata also evaluates when structures should and should not exhibit join relationships. 

    Within the imperfect world of data, analysts may be unknowingly confronted with structures that should join but don’t (or vice versa).  This can adversely affect dependent analyses.

Automated Data Quality with Live Metadata

Drawing from the rich insights captured within its live metadata repository, Lightwell automates the construction of comprehensive data validation architecture. 

All of the defined content and applicability standards are combined into comprehensive validation trees, against which your company’s data can be evaluated. 

The results of these evaluations (and the corresponding corrective actions) are written to a central, query-friendly log to simplify the monitoring of data quality.

Let’s Get Started

Are you ready to truly know your data?   Get in touch with us and let’s review how the Lightwell metadata repository and data quality architecture can improve your analytics.

Contact us today or call us at +1 (614) 310-2700, and we’ll connect you with one of our experts.