DATA QUALITY

Opportunities for improving data quality

Data Quality ensures the pre-treatment of data that allows the optimization of its use, either by means of correction or by reorganizing it.

Quality in data, quality in information

Data normalization, according to the requirements established by users, is an increasingly critical activity for business management. The existence of a data quality assessment process is essential both to allow cross-checking of different sources (standardization of expressions) and to ensure that analyzes and segmentations are rigorous (identification of invalid or duplicates). Even in a simple database, a process of this type makes it possible to find duplicate records, information gaps or inconsistencies. They may seem like details, but they can have a big impact on the final result. In addition, preparatory work such as standardization of expressions or ordering can have a major impact on subsequent operational processes (for example for the purpose of optimizing shipping costs when mail is delivered to courier operators in a certain way).

When starting a Data Quality process we follow the following steps:

R

Requirements gathering:

The purpose of the analysis or process must be clear in order to define the requirements that the data must meet.

R

Data profile identification:

We carefully examine the following aspects of the data: format, standards, record consistency, value distributions and outliers and whether the records are complete.

R

Identification of associated flows:

What they are used for, what tools, what intersections will exist, where they are available for updating and by whom and what processes are associated with the update.

R

Implementation:

In addition to creating rules and validations in accordance with the information collected in the previous steps, we ensure:

$

Integrity: The use of primary and foreign keys assumes a crucial role in the case of relational bases, in situations of multiple unrelated systems we ensure the existence of validation of additional conditions to the format (check constraint) and the use of mechanisms triggered by specific actions (triggers).

$

Traceability: Whenever a problem is detected in a log, we guarantee that it is possible to quickly identify its origin and correct it.

$

Completeness: If necessary, the process may include crossing with external data sources, enriching the data and generating information with great potential..