Toward Data-Driven Requirements Engineering

Developers, requirements analysts, and managers could systematically use explicit and implicit user feedback in an aggregated form to support requirements decisions. The goal is data-driven requirements engineering by the masses and for the masses. Analytics tools can help classify and filter user feedback according to the information it contains. Researchers have studied how to use text classification, natural-language processing, and other to categorize reviews. Automatic feedback classification can provide an overall picture of app usage and user engagement. This can also be used for comparing releases over time (in terms of the provided and requested requirements or the bug reports) or comparing similar apps. Much research exists on collecting and processing usage data for software engineering, focusing mainly on error reproduction and localization. The shift from reactive to real-time and even proactive decision making is crucial in developing software-intensive products. To create products rapidly with higher customer acceptance, developers must incrementally build and deploy products that rely on deep customer insight and real-time feedback. Changing software engineering teams? mind-set to accept users as equal stakeholders with potentially good ideas and suggestions is an important cultural challenge.