The rapid advancement of technology has made tablets indispensable in both personal and professional settings. However, despite their widespread use, understanding the features that drive customer satisfaction and areas needing improvement remains a critical challenge. This thesis aims to identify and analyze the key factors influencing customer sentiments towards tablet features.

The study utilized advanced sentiment analysis techniques, including VADER, LDA, and BERTopic, to process and analyze a comprehensive dataset of customer reviews. The textual data was transformed into high-dimensional vectors using the BAAI/bge-m3 embedding model, which were then reduced and clustered using UMAP and HDBSCAN for efficient analysis.

The findings reveal that features such as size, brightness, and value for money are highly appreciated by customers, as indicated by positive sentiment scores. On the other hand, battery life and charging time were identified as significant pain points, receiving notably lower sentiment scores. These results highlight the variability in tablet performance and quality across different models and underscore the need for targeted improvements.

In conclusion, this research provides valuable insights for tablet manufacturers seeking to enhance product features in line with customer expectations. Manufacturers can enhance customer satisfaction and gain a competitive edge by addressing identified areas for improvement. Additionally, employing sophisticated sentiment analysis tools is crucial for gaining a comprehensive understanding of customer experiences.