Ever felt annoyed when an app on your phone suddenly crashes or slows down? Those frustrations, often shared in user reviews, actually hold valuable insights. A recent article titled “Topic Modeling for User Feedback Dataset” published in Communications in Mathematical Biology and Neuroscience (2025) shows how app users’ feedback can be transformed into actionable knowledge to build better digital services.
The study was conducted by Sinta Septi Pangastuti, Eneng Nunuz Rohmatullayaly, and Nuroh Najmi from the Faculty of Mathematics and Natural Sciences, Universitas Padjadjaran. They analyzed more than 15,000 user feedback entries from 15 different mobile applications across various categories. Instead of relying on older methods like Latent Dirichlet Allocation (LDA), the team applied Top2Vec, a modern topic modeling technique that leverages word embeddings and clustering algorithms to uncover semantic patterns in text.
The results are impressive. Top2Vec can automatically determine the number of topics and generate richer, more interpretable themes—from technical bugs and performance issues to user experience concerns. Evaluation using Coherence Score and Topic Diversity confirmed that this method outperforms traditional approaches.
This research also contributes to the achievement of several Sustainable Development Goals (SDGs):
- SDG 9 (Industry, Innovation, and Infrastructure): fostering digital innovation through data-driven analysis.
- SDG 12 (Responsible Consumption and Production): helping developers create more reliable and responsive applications.
- SDG 16 (Peace, Justice, and Strong Institutions): improving transparency and trust by enhancing digital services.
In other words, what often seems like mere complaints can become the key to building smarter, more sustainable, and user-friendly apps—thanks to the power of artificial intelligence and statistical innovation.
Source: https://www.scopus.com/record/display.url?eid=2-s2.0-85217213519&origin=resultslist
07/Stat/2025
