Requirements Behind Reviews: How do Software Practitioners See App User Reviews to Think of Requirements?

• Walter Takashi Nakamura UFAM
• José Cezar de Souza Filho UFAM
• Lígia Márcia Teixeira UFAM
• Afonso Silva UFAM
• Rogenis da Silva UFAM
• Bruno Gadelha UFAM
• Tayana Conte UFAM

Resumo

The popularity of mobile computing has led to the development of several mobile apps to solve end-users everyday problems. Nowadays, app stores allow users to write reviews reporting their experiences using the apps. Negative experiences can reveal the reasons why users are moving from one app to another and drive app adoption by future users. Thus, software companies should listen to users’ voices to meet their needs and improve the quality of their apps. In this context, Data-Driven Requirements Engineering (DDRE) emerges as an alternative approach to develop and manage requirements from large volumes of data such as user reviews. Most research on DDRE has focused on assessing different machine learning, information retrieval, and natural language processing techniques for automatic requirements mining and classification. However, we need to move forward to understand how to support the DDRE adoption by software practitioners who will apply the paradigm in practice. In this paper, we present a tool-supported approach for data-driven requirements elicitation. It applies text mining techniques to obtain user reviews from app stores and presents the most frequent terms with their associated reviews through a visual representation. To advance the understanding of DDRE adoption, we conducted a feasibility study with industry practitioners to investigate how our approach supports the requirements elicitation process. Our main findings indicated a positive acceptance by practitioners. They perceived it as easy to use and useful to think of requirements while demonstrating an intention to adopt it in the requirements elicitation process.

Palavras-chave: data-driven requirements engineering, app user reviews, requirements elicitation