Abstract
This study developed a hybrid Decision Support System (DSS) that combines descriptive data analytics with decision-making knowledge. The system was designed to assist decision-makers in a harness manufacturing company in South Africa, aiming to improve decision-making in quality management and reduce the recurrence of quality issues.
The study aimed to provide practical, real-world solutions for Company X and was grounded in a pragmatist research philosophy. Pragmatism prioritises actionable outcomes and emphasises using methods best suited to addressing the research problem. A deductive approach was employed to meet the study’s objectives. This approach facilitated the application of established theories and frameworks, while also allowing unique insights to emerge from the analysis of the data from Company X. Moreover, a mixed-methods design was adopted, combining qualitative and quantitative data collection and analysis to ensure a holistic understanding of the organisation’s challenges and decision-making processes.
A case study research strategy was selected to enable an in-depth exploration of the unique context of Company X. This strategy was critical in ensuring that the developed DSS was tailored to the organisation’s specific challenges and operational environment. Data collection followed a cross-sectional time horizon, capturing a snapshot of the organisation’s challenges and resources at a single point in time. This approach allowed the study to deliver a practical, functional, and timely solution aligned with the needs of the organisation.
The research adhered to ethical guidelines before data collection, ensuring that all data was collected only after approval from the relevant ethics committee. Company X's confidentiality was safeguarded to protect its reputation and operational integrity. Data collection was transparent, with stakeholders informed about the study's objectives and scope. Data was securely stored, access was restricted, and applicable data protection regulations were adhered to. Informed consent was obtained from all participants, ensuring their understanding and agreement to partake in the study.
Secondary quantitative data on reworked and scrapped harnesses from 2019 to 2023 were collected from an existing database in the form of Excel sheets. The study’s population focused on non-conforming harnesses recorded during this period, and a census sampling method was used to allow for in-depth analysis. In addition, qualitative data were collected from the latest
v
work instructions that guide operators on whether to rework or scrap defective products. Only valid work instructions relevant to handling non-conformances were selected and analysed using purposive sampling. Systematic analysis of the quantitative data was conducted to identify trends and patterns, while thematic analysis of the work instructions, performed using Atlas.ti version 23, revealed the criteria and processes used in decision-making.
Findings from the systematic analysis showed that data-driven decision-making in Company X heavily relied on Excel, focusing on weekly rework and scrap frequencies and the reasons for scrapping or reworking parts. The thematic analysis highlighted that all defective parts could be reworked except under specific conditions, such as defects in safety-critical areas, defects on wires passing through grommets, or faults in certain harness locations.
Using these insights, a tailored DSS was developed using Python and Streamlit to address the limitations of the current decision-making process. The new system offered significant improvements, including enhanced data insights, a structured database, and a user-friendly interface. These capabilities enable more effective management of non-conformances, providing Company X with a robust tool to overcome its quality management challenges.