Data Quality Assessment

Data quality standards are usually created based on the perspective of data producers, who historically had been the data consumers. However, in the modern world, data users frequently are not data producers, which makes data quality control more difficult. Following this trend, the authors in Ref. [18] used the traditional data quality dimensions and defined their elements and indicators under the view of BD requirements.

The most widely accepted data quality dimensions are as follows:

Availability - determines how affordable and convenient users can collect data and its associated information. It is described by three elements: accessibility, authorization, and timeless.

Usability - refers to the capacity of the data to fulfill users' objectives in terms of definition/documentation, reliability, and metadata. Reliability -this is the data quality dimension that indicates how reliable the data are and their elements such as accuracy, consistency, completeness, integrity, and audibility.

Relevance - expresses the level of correlation between data content and consumers' desires and requirements.

Presentation quality - refers to a proper representation of the data to make users understand the data.

The methodology was complemented by the assignment of 1-5 elements for each dimension to create a Data Quality assessment framework, as demonstrated in Table 3.1.

Some dimensions and elements can be measured objectively (completeness, timeliness, integrity), and others rely heavily on context or on subjective interpretation (usability, relevance, authorization).

As previously discussed, for each specific business problem and environment, the choice of data quality elements to be used will differ. For each selected dimension, there will be different tools, methods, and processes, and consequently, variations in the required time, costs, and workforce. The first step of the assessment cycle is the definition of the data collection objectives, which are based on business strategy and requirements - for example, operations, decision-making, and planning [18].

 
Source
< Prev   CONTENTS   Source   Next >