Data Analytics Applications: How to maintain and keep them fit for purpose

Abstract

Data-driven decision making is critical for organizational success, yet widely adopted methodologies like CRISP-DM lack comprehensive guidance for maintaining Data Analytics Applications (DAAs) after deployment. This research investigates how organizations keep their DAAs up-to-date and suitable for their intended purpose through a qualitative case study analysis of maintenance practices across two organizations. The study reveals that organizations have developed practical maintenance approaches without systematic structure advocated in academic literature, primarily adopting Agile frameworks, particularly SAFe, rather than specialized data mining process models. Organizations allocate 20-25% of their DAA development capacity to maintenance activities, demonstrating partial alignment with theoretical frameworks emphasizing continuous monitoring. However, gaps do exist between theoretical recommendations and practical implementations, with organizations operating reactively rather than proactively regarding DAA maintenance. Key challenges include knowledge retention problems, data quality issues, system complexity and governance gaps, and emerging compliance requirements. The findings suggest that existing theoretical frameworks require enhancement to address practical implementation needs. The research recommends integrating data analytics-specific requirements within established Agile frameworks, developing systematic maintenance approaches grounded in recognized frameworks, and formalizing resource allocation for maintenance activities. This study contributes to bridging the gap between academic theory and industry practice in DAA maintenance.

Date of Award	22 Jun 2025
Original language	English
Supervisor	Remko Helms (Examiner) & Khoi Nguyen (Co-assessor)