CoFFeR: A Unified Dataset of Commits, Feedback, Feature Descriptions and Release Notes

Summary of ArtefactsThese artefacts accompany the publication "CoFFeR: A Release-Related Dataset of Commits, Feedback, Feature Description, Release Notes" at the 23rd {IEEE/ACM} International Conference on Mining Software Repositories, MSR@ICSE 2025, Rio de Janeiro, Brazil. The dataset includes several artefacts, surrounding software releases.

Author and Article DetailsVolodymyr Biryuk (Universität Hamburg) - [e-mail]

Jialiang Wei (Universität Hamburg) - [e-mail]

Prof. Dr. Walid Maalej (Universität Hamburg) - [e-mail]

Please cite this work as:

Title of Publication: "CoFFeR: A Release-Related Dataset of Commits, Feedback, Feature Description, Release Notes"

Abstract of the Chapter:

User feedback plays a vital role in software and requirements en- gineering, serving as a bridge between user needs and developer actions. Understanding the relationship between user feedback and app releases is essential, as it reveals how developers respond to user concerns through updates, bug fixes, or new features. However, it remains unclear how user feedback actually influences develop- ers’ decisions and the resulting software evolution. To address this gap, we present a novel dataset that connects user feedback to software releases for five open-source mobile apps. Unlike previ- ous datasets which limit their focus on Google Play or GitHub, our dataset integrates additional information from diverse sources, including community forums, issue tracking systems and release notes. Specifically, it encompasses ≈1M pieces of user feedback from Google Play, community forums and issue trackers, ≈700 release notes from company websites and GitHub, ≈86,000 code changes from GitHub, and feature descriptions from the according documentation websites of the apps. This dataset provides a foun- dation for further empirical studies on the impact of user feedback in software evolution and enables researchers and practitioners to analyze patterns in feedback-driven development.

Description of ArtefactsThis package is split into following parts:

data: The directory containing the data described in the publication.github: The directory to save all scraped GitHub repositories.mongodb.zip: The final data of this publication.src: The directory containing the replication code to recreate the data.scraping: This directory contains code to download all data.data: This directory contains JupyterNotebooks that visualize data stats.LicencesThe code is licensed under MIT. The licence is included in this repository and further information can be found here: https://opensource.org/licenses/MIT

The data is licensed under CC BY 4.0. The licence is included in this repository and further information can be found here: https://creativecommons.org/licenses/by/4.0/

Resources