Sentiment Analysis of Google Play and App Store Reviews of Threads

This research explores user sentiment of Meta's social networking app Threads, which is a part of another social networking application, Instagram, through sentiment analysis of its reviews on Google Play Store and App Store. The research uses a multi - pronged strategy that combines Kmeans clustering for spam identification, sentiment analysis, and keyword-specific sentiment analysis, using python as the programming language to find output. K-means clustering is first applied to remove spam from user reviews. The non-spam reviews are next subjected to sentiment analysis using RoBERTa – base, a natural language processing model,. Owing to its strong architecture, RoBERTa - base can process reviews with emojis and text in several languages with efficiency, guaranteeing accuracy sentiment analysis across a wide range of user-generated content. Furthermore, in order to identify the sentiment patterns related to a few important keywords, sentiment analysis is carried out independently on reviews that contain important keywords like “privacy” and “notifications.” A word cloud is created for all the non-spam reviews. Word clouds are also created for each category after the reviews are divided into groups according to the overall positive and negative sentiment scores in order to better investigate the opinions expressed. We will also generate a sentiment distribution of reviews over a timeline for better understanding of the target consumer's sentiments. Through the integration of sophisticated sentiment analysis techniques like RoBERTa - base, this study offers profound insights into users' perceptions and opinions of Threads, illuminating areas of user satisfaction and concern with the application.