Requirements engineering for data analytics projects: A practical framework for early-stage planning

Abstract

Data analytics projects depend not only on analytical techniques but also on the quality of the requirements defined before data acquisition and analysis begin. Unclear goals, poorly specified data sources, and mismatched expectations about outputs can increase project complexity and undermine the usefulness of results. This paper examines how requirements engineering can support the early planning of data analytics projects. It synthesises literature on data analytics, data-driven decision-making, and requirements engineering, with particular attention to the generic requirements template proposed by Kumarasinghe and Kirikova for descriptive and diagnostic analytics. The paper organises key requirements around four phases – initiation, acquisition, analysis, and presentation – and develops a practical checklist for discussions between analysts and stakeholders. The checklist covers project objectives, data sources and properties, data preparation, analytical methods, tools, result validation, and reporting requirements. It is intended to improve completeness and communication at the project’s outset, while recognising that detailed requirements must be refined for each project context.