Haluk Demirhan

Assessing Data Quality Management Issues in Open-Sourced Data Sets with Statistical Methodology

Haluk Demirhan · MSc in Management of Business Information Technology dissertation, University of Greenwich · Submitted 20 December 2024 · Supervisor: Dr Solomon Ebenuwa

Download the dissertation (PDF, 54 pages)

Abstract

Open-source datasets are critical resources in the data-driven era, driving innovation across fields like artificial intelligence, machine learning, and business intelligence. However, these datasets often suffer from inconsistent quality, hindering their effective application. This dissertation addresses this challenge by developing a comprehensive R package, DataQualityPackage, for assessing and improving the quality of open-source datasets. The framework evaluates key data quality dimensions, including completeness, accuracy, consistency, and timeliness, while incorporating ethical considerations such as data privacy and bias detection. A design science methodology guided the package's development, including features like missing data imputation, outlier detection, correlation analysis, and distribution assessment. The package's effectiveness was validated using datasets from platforms like Kaggle, demonstrating its capability to identify and mitigate quality issues. Visual and statistical results confirm the framework's utility in enhancing data reliability.

This research provides a standardised tool for data quality assessment that contributes to academic and industry practices. By addressing existing gaps, the framework sets a foundation for scalable and ethical data validation methods. Future work will focus on expanding the package's features and exploring its integration with advanced data analytics pipelines.

Keywords: open-source data quality; data validation framework; R programming; data completeness; data accuracy; data consistency; data timeliness; outlier detection; missing data imputation.

Software: datagrade on CRAN

The method is released as the R package datagrade, published on CRAN (version 0.2.0, 29 October 2026). It scores a data set on the ISO/IEC 25012 data quality model. Install it in R with install.packages("datagrade").

This page holds the dissertation as submitted in December 2024. datagrade is a later, reworked implementation of its method.

Cite as

Demirhan, H. (2024). Assessing Data Quality Management Issues in Open-Sourced Data Sets with Statistical Methodology. MSc dissertation, University of Greenwich.

@mastersthesis{demirhan2024dataquality,
  author = {Demirhan, Haluk},
  title  = {Assessing Data Quality Management Issues in Open-Sourced Data Sets with Statistical Methodology},
  school = {University of Greenwich},
  year   = {2024},
  type   = {{MSc} dissertation},
  url    = {https://halukdemirhan.com/thesis/}
}