Prioritizing User Feedback from Twitter: A Survey Report

Abstract

Twitter messages (tweets) contain important information for software and requirements evolution, such as feature requests, bug reports and feature shortcoming descriptions. For this reason, Twitter is an important source for crowd-based requirements engineering and software evolution. However, a manual analysis of this information is unfeasible due to the large number of tweets, its unstructured nature and varying quality. Therefore, automatic analysis techniques are needed for, e.g., summarizing, classifying and prioritizing tweets. In this work we present a survey with 84 software engineering practitioners and researchers that studies the tweet attributes that are most telling of tweet priority when performing software evolution tasks. We believe that our results can be used to implement mechanisms for prioritizing user feedback with social components. Thus, it can be helpful for enhancing crowd-based requirements engineering and software evolution.