[Paper Review] Exploring Data Management Challenges and Solutions in Agile Software Development: A Literature Review and Practitioner Survey

Ahmed Fawzy, Amjed Tahir+2 more|arXiv (Cornell University)|Feb 1, 2024

Big Data and Business Intelligence5 citations

TL;DR

This paper conducts a systematic literature review to identify data management challenges in agile software development and surveys proposed solutions, outlining their implications for practice.

ABSTRACT

Context: Managing data related to a software product and its development poses significant challenges for software projects and agile development teams. These include integrating data from diverse sources and ensuring data quality amidst continuous change and adaptation. Objective: The paper systematically explores data management challenges and potential solutions in agile projects, aiming to provide insights into data management challenges and solutions for both researchers and practitioners. Method: We employed a mixed-methods approach, including a systematic literature review (SLR) to understand the state-of-research followed by a survey with practitioners to reflect on the state-of-practice. The SLR reviewed 45 studies, identifying and categorizing data management aspects along with their associated challenges and solutions. The practitioner survey captured practical experiences and solutions from 32 industry practitioners who were significantly involved in data management to complement the findings from the SLR. Results: Our findings identified major data management challenges in practice, such as managing data integration processes, capturing diverse data, automating data collection, and meeting real-time analysis requirements. To address the challenges, solutions such as automation tools, decentralized data management practices, and ontology-based approaches have been identified. The solutions enhance data integration, improve data quality, and enable real-time decision-making by providing flexible frameworks tailored to agile project needs. Conclusion: The study pinpointed significant challenges and actionable solutions in data management for agile software development. Our findings provide practical implications for practitioners and researchers, emphasizing the development of effective data management practices and tools to address those challenges and improve project success.

Motivation & Objective

• Identify data management aspects discussed in agile software development.
• Characterize challenges across data management aspects (e.g., integration, collection, quality, analysis).
• Identify and categorize proposed solutions to data management challenges in agile contexts.
• Assess implications of data management challenges on teams and product delivery.
• Provide recommendations for integrating robust data management into agile frameworks.

Proposed method

• Followed Kitchenham and Charters SLR guidelines.
• Used Scopus to search for English-language studies before Oct 2023.
• Applied inclusion/exclusion criteria to filter studies.
• Retrieved 45 high-quality studies through a two-stage screening and full-text review.
• Extracted data and categorized studies by data management aspects.
• Provided replication dataset with extracted results.

Figure 1: The review process

Experimental results

Research questions

• RQ1 RQ1: What data management aspects are discussed in the context of agile software development?
• RQ2 RQ2: What challenges are faced regarding these aspects?
• RQ3 RQ3: What solutions are proposed to address these challenges?

Key findings

• Fifteen data management aspects were identified; data integration, data collection, data quality, and data analysis are among the most discussed.
• Major challenges include interoperability, semantic heterogeneity, and integrating heterogeneous data sources.
• Solutions include ontologies, data mesh approaches, automated ETL tools, and quality-focused development methods.
• Cloud-based data-loading pipelines, ontology-based integration, and architecture-centric approaches were among the implemented solutions.
• There is evidence of overlap between data collection and analysis, as well as between data integration and quality.

Figure 2: Total number of studies discussing challenges and solutions associated with fifteen aspects of data management