An Approach to Data Quality Evaluation

Analysis of Open Health Data Quality Using Data Object-Driven Approach to Data Quality Evaluation: Insights From a Latvian Context

“…As a result, a simple and constructive approach to data quality evaluation is needed in order to provide users with possibility to take part in open data analysis even if they don't have deep knowledge in IT and data quality areas. The data object-driven approach to data quality evaluation previously proposed in (Bicevskis, 2018a;Nikiforova, 2019) is considered as highly appropriate for given purpose. It doesn't use "quality dimension" concept, that became traditional for data quality solutions, as there is no universal classification that should be followed by data quality solutions' developers and users, despite the high number of attempts.…”

“…Taking into account the common data quality division in syntactic and semantic accuracy (Scannapieco, et al, 2005) as well as multiple issues appearing from "quality dimension" concept, the proposed approach avoids it, respecting wider concept "data quality requirement" where "data quality dimension" is its subset. Syntactic accuracy of the proposed solution can be found in (Bicevskis et al, 2018a;Nikiforova, 2018b), however the semantic accuracy was presented in (Nikiforova and Bicevskis, 2019). As the comprehensive discussion on the existing solutions, rationale and detail of the given approach were discussed previously, they won't be discussed in detail here, explaining only the basics of the solution that should be understood by the users.…”

“…In addition, as open data are usually used by wide audience that may not have deep knowledge in IT or data quality areas, a solution that should be applied to open data sets should be simple enough ensuring particular users with possibility to take part in the analysis of open data. To achieve this aim previously proposed user-oriented data object-driven approach (Bicevskis, 2018a;Nikiforova, 2019) should be examined in order to ensure it is applicable on such data sets providing all necessary functionality making its quality analysis.…”

Towards Data Quality Runtime Verification

“…It is one of the main benefits of this approach as usually approaches for data quality evaluation are suitable mostly for IT-and DQ-experts requiring deep knowledges in both areas [6]. However, data object-driven model can be used by wide audience without need to make in-depth analysis of the basics of the approach as it usually happens with other approaches where at least an exploration of the list of dimensions, their meanings and criteria under each of them, needs a lot of time for every particular solution as criteria differentiate from case to case [3], [6], [7]. At the same time, the proposed approach already demonstrated its effectiveness by applying it to real datasets [4], [6] - [8].…”

“…However, runtime verification of business processes checks only the correctness of process execution based on the evaluation of process execution sequence, meanwhile this research proposes to extend it by involving data quality mechanism, that will assure the specific process did not negatively affect data stored in information systems that were affected by this process. It is achieved by applying data object-driven approach to data quality evaluation [3]. This approach is based on definition of data object which quality should be analysed, quality requirements definition that are applied to the parameters of the defined data object and measuring data quality.…”