Emitza Guzman

How Do Users Like This Feature? A Fine Grained Sentiment Analysis of App Reviews

App stores allow users to submit feedback for downloaded apps in form of star ratings and text reviews. Recent studies analyzed this feedback and found that it includes information useful for app developers, such as user requirements, ideas for improvements, user sentiments about specific features, and descriptions of experiences with these features. However, for many apps, the amount of reviews is too large to be processed manually and their quality varies largely. The star ratings are given to the whole app and developers do not have a mean to analyze the feedback for the single features. In this paper we propose an automated approach that helps developers filter, aggregate, and analyze user reviews. We use natural language processing techniques to identify fine-grained app features in the reviews. We then extract the user sentiments about the identified features and give them a general score across all reviews. Finally, we use topic modeling techniques to group fine-grained features into more meaningful high-level features. We evaluated our approach with 7 apps from the Apple App Store and Google Play Store and compared its results with a manually, peer-conducted analysis of the reviews. On average, our approach has a precision of 0.59 and a recall of 0.51. The extracted features were coherent and relevant to requirements evolution tasks. Our approach can help app developers to systematically analyze user opinions about single features and filter irrelevant reviews.

How can i improve my app? Classifying user reviews for software maintenance and evolution

App Stores, such as Google Play or the Apple Store, allow users to provide feedback on apps by posting review comments and giving star ratings. These platforms constitute a useful electronic mean in which application developers and users can productively exchange information about apps. Previous research showed that users feedback contains usage scenarios, bug reports and feature requests, that can help app developers to accomplish software maintenance and evolution tasks. However, in the case of the most popular apps, the large amount of received feedback, its unstructured nature and varying quality can make the identification of useful user feedback a very challenging task. In this paper we present a taxonomy to classify app reviews into categories relevant to software maintenance and evolution, as well as an approach that merges three techniques: (1) Natural Language Processing, (2) Text Analysis and (3) Sentiment Analysis to automatically classify app reviews into the proposed categories. We show that the combined use of these techniques allows to achieve better results (a precision of 75% and a recall of 74%) than results obtained using each technique individually (precision of 70% and a recall of 67%).

Sentiment analysis of commit comments in GitHub: an empirical study

Emotions have a high impact in productivity, task quality, creativity, group rapport and job satisfaction. In this work we use lexical sentiment analysis to study emotions expressed in commit comments of different open source projects and analyze their relationship with different factors such as used programming language, time and day of the week in which the commit was made, team distribution and project approval. Our results show that projects developed in Java tend to have more negative commit comments, and that projects that have more distributed teams tend to have a higher positive polarity in their emotional content. Additionally, we found that commit comments written on Mondays tend to a more negative emotion. While our results need to be confirmed by a more representative sample they are an initial step into the study of emotions and related factors in open source projects.

Retrieving Diverse Opinions from App Reviews

Context: Users can have conflicting opinions and different experiences when using software and user reviews serve as a channel in which users can document their opinions and experiences. To develop and evolve software that is usable and relevant for a diverse group of users, different opinions and experiences need to be taken into account. Goal: In this paper we present DIVERSE, a feature and sentiment centric retrieval approach which automatically provides developers with a diverse sample of user reviews that is representative of the different opinions and experiences mentioned in the whole set of reviews. Results: We evaluated the diversity retrieval performance of our approach on reviews from seven apps from two different app stores. We compared the reviews retrieved by DIVERSE with a feature-based retrieval approach and found that on average DIVERSE outperforms the baseline approach. Additionally, a controlled experiment revealed that DIVERSE can help develop- ers save time when analyzing user reviews and was considered useful for detecting conflicting opinions and software evolution. Conclusions: DIVERSE can therefore help developers collect a comprehensive set of reviews and aid in the detection of conflicting opinions.

Visualizing emotions in software development projects

Developers and managers need to be aware of the emotional climate of the projects they are involved to take corrective actions when necessary and to have a better understanding of the social factors affecting the project. With the growing trend of distributed teams and textual communication this type of awareness is more difficult to obtain and maintain. We propose to improve emotional climate awareness in software development projects by means of a visualization prototype which includes general and detailed views of the topics and emotions expressed in software project collaboration artifacts. We performed an initial case study in which the mailing list content of a software project was visualized. The study suggests that the length, frequency and emotion diversity of the exchanged content varies according to the project phase. However, a more extensive evaluation needs to be made.

Automated Requirements Extraction for Scientific Software

Abstract Requirements engineering is crucial for software projects, but formal requirements engineering is often ignored in scientific software projects. Scientists do not often see the benefit of di- recting their time and effort towards documenting requirements. Additionally, there is a lack of requirements engineering knowledge amongst scientists who develop software. We aim at helping scientists to easily recover and reuse requirements without acquiring prior requirements engineering knowledge. We apply an automated approach to extract requirements for scientific software from available knowledge sources, such as user manuals and project reports. The ap- proach employs natural language processing techniques to match defined patterns in input text. We have evaluated the approach in three different scientific domains, namely seismology, build- ing performance and computational fluid dynamics. The evaluation results show that 78–97% of the extracted requirement candidates are correctly extracted as early requirements.

FAVe: Visualizing User Feedback for Software Evolution

App users can submit feedback about downloaded apps by writing review comments and giving star ratings directly in the distribution platforms. Previous research has shown that this type of feedback contains important information for software evolution. However, in the case of the most popular apps, the amount of received feedback and its unstructured nature can produce difficulties in its analysis. We present an interactive user feedback visualization which displays app reviews from four different points of view: general, review based, feature based and topic-feature based. We conducted a study which visualized 2009 reviews from the Dropbox app available in the App Store. Participants considered the approach useful for software evolution tasks as they found it could aid developers and analysts get an overview of the most and least popular app features, and to prioritize their work. While using different strategies to find relevant information during the study, most participants came to the same conclusions regarding the user reviews and assigned tasks.