ARCHIVES

The rapid growth of music streaming platforms like Spotify has revolutionized music consumption, generating an exponential increase in user-generated textual data through app reviews. Analysing such large volumes of unstructured text manually to understand user behaviour is impractical and time-consuming. Sentiment analysis, a key application of Natural Language Processing (NLP), presents a valuable tool for automatically identifying the emotional tone expressed in textual data, gauging user satisfaction, and pinpointing areas for improvement. The study utilizes a Spotify user review dataset comprising 61,594 collected reviews. To ensure high-quality input for the model, extensive data preprocessing techniques were applied using the NLTK library. The cleaned textual data, stored as refined review descriptions, was then structured so the neural network could effectively process the vocabulary and underlying sentiments. A deep learning neural network architecture was proposed and implemented as the core model for this sentiment classification task. To prevent over fitting and ensure the model learns generalized patterns, the network was trained iteratively over 5 epochs to optimize its ability to distinguish between different sentiments polarities. The performance of the proposed model was evaluated using accuracy and loss metrics during the training and validation phases. The experimental results demonstrate that the neural network achieved an outstanding training accuracy of 94.66% and a validation accuracy of 93.20%, with a final validation loss of 0.1427. The findings of this research confirm that deep learning approaches provide highly accurate and effective sentiment analysis applications in the music streaming industry.