Difference between revisions of "Data Quality"

Latest revision as of 23:01, 27 September 2021

Definition

Data Quality (also Data Integrity) refers to the condition of information sets (data) that are to be used as inputs for qualitative or quantitative risk assessment e.g. in the form of Portfolio Information, Algorithms and/or other decision support tools

Regulated institutions are required to have in place a formal Data Quality Management Framework^[1]

Examples

In order to support to the development of an internal Credit Scorecard, a firm must have access to historical credit data that meet data quality criteria
In Data Privacy context data quality^[2] refers to a set of principles laid down in Article 5 of the GDPR and Article 4 of Regulation (EU) 2018/1725, namely:
- Lawfulness, fairness and transparency
- Purpose limitation
- Data minimisation
- Accuracy
- Storage limitation
- Integrity and confidentiality

Issues and Challenges

During the Financial Crisis data quality was identified as a contributing cause to poor risk management ^[3]
Data quality issues are a significant component of Model Risk, colloquially referred to as the "garbage in, garbage out" principle.

References

↑ ECB guide to internal models - Credit Risk, Sep 2018
↑ EDPS Glossary
↑ BCBS 239: Principles for effective risk data aggregation and risk reporting

[1] ECB guide to internal models - Credit Risk, Sep 2018

[2] EDPS Glossary

[3] BCBS 239: Principles for effective risk data aggregation and risk reporting

[1]

[2]

[3]

@@ Line 4: / Line 4: @@
 Regulated institutions are required to have in place a formal [[Data Quality Management Framework]]<ref>ECB guide to internal models - Credit Risk, Sep 2018</ref>
-=== Data Quality Assurance ===
+== Examples ==
-Data quality assurance is a planned and systematic set of processes aiming to provide the desired confidence that the information embodied in a given data set conforms to established requirements. Data quality considerations are typically grouped as follows:
+* In order to support to the development of an internal [[Credit Scorecard]], a firm must have access to historical credit data that meet data quality criteria
+* In [[Data Privacy]] context data quality<ref>[https://edps.europa.eu/data-protection/data-protection/glossary/ EDPS Glossary]
-* Data Quality Assessment. Data quality assessment is a highly contextual process (dependent on the intended uses of the data) that establishes metrics of data quality along a number of different dimensions ([[Data Quality Standards]])
+</ref> refers to a set of principles laid down in Article 5 of the GDPR and Article 4 of Regulation (EU) 2018/1725, namely:
-* [[Data Integrity Validation | Data Validation]] that is primarily focused on validating the integrity of data
+** Lawfulness, fairness and transparency
-* [[Data Cleansing]], the process of correcting and possibly transforming data in order to produce a set that is suitable for use
+** Purpose limitation
+** Data minimisation
-== Example ==
+** Accuracy
-In order to support to the development of an internal [[Credit Scorecard]], a firm must have access to historical credit data that meet data quality criteria
+** Storage limitation
+** Integrity and confidentiality
 == Issues and Challenges ==
@@ Line 25: / Line 26: @@
 [[Category:Data Quality]]
+[[Category:Data Privacy]]
 [[Category:ECB TRIM]]
 [[Category:Credit Scoring Business Requirements]]