(TFM) RE-Miner 2.0: A Holistic Web Application for Extensive Review-Based Analysis

Description: This thesis aims to develop RE-Miner 2.0, an enhanced version of the existing tool. The project will extend RE-Miner's capabilities by integrating additional metrics for analyzing user reviews, focusing on categorizing review types, identifying discussed topics, and assessing the emotional tone and rating patterns within reviews. These additions will allow for a more comprehensive analysis of user feedback, enabling stakeholders to derive actionable insights across diverse review dimensions.

The project will involve redesigning RE-Miner's architecture to support the integration of new analysis pipelines while maintaining modularity and scalability. Advanced natural language processing (NLP) techniques will be employed for tasks such as classification, sentiment analysis, and topic modeling, coupled with interactive visualizations to present the results in an intuitive and accessible manner.

Objectives:

Extend RE-Miner's architecture to support a broader range of analytical metrics.
Develop and integrate advanced NLP pipelines for deeper review analysis.
Enhance data visualization capabilities to present results in a user-friendly and actionable format.

Expected Outcome:
An upgraded tool that provides a holistic platform for app review analysis, empowering developers, product managers, and stakeholders to derive insights critical to requirements engineering and product development.

Degree: MDS/MEI

Research Areas: Requirements Engineering, User Feedback Analysis, Natural Language Processing, Sentiment Analysis, Web-based Application Development

Technologies: Python, Django/Flask, NLP Libraries (e.g., Hugging Face, SpaCy), Topic Modeling Tools (e.g., BERTopic, LDA), SQL/NoSQL Databases, Data Visualization Frameworks (e.g., D3.js, Plotly), React.js/Angular for front-end, Docker for deployment

Contact: Quim Motger

e-mail: joaquim.motger (at) upc.edu