Unveiling Emotional Landscape: A Sentiment Expedition into Instagram Play Store and App Store Reviews using TextBlob

Abstract

Using natural language processing techniques applied to user evaluations from the Google Play Store and Apple App Store, this article analyzes and predicts user sentiment toward Instagram. Understanding user sentiment trends, visualizing changes in user satisfaction, and forecasting future sentiment patterns are the main goals, providing app developers with insightful information. Our approach is divided into three phases: future sentiment predictions, periodic trend visualization, and sentiment extraction. We utilize TextBlob, a Python-based natural language processing package that can handle important tasks including sentiment analysis, tokenization, lemmatization, and part-of-speech tagging, for sentiment extraction. We provide an immediate overview of user viewpoints by using TextBlob to record and classify user input into positive, negative, and neutral attitudes. To see trends and spot notable shifts in user satisfaction, we then create a time-phase model that shows sentiment differences over a monthly period. We apply an Auto-Regressive Integrated Moving Average model from the statsmodel package to forecast sentiment trends over the next two months by analyzing historical sentiment data. According to experimental results, our model predicts user attitudes toward the Instagram program with high reliability, achieving an accuracy of 93.27% when comparing predicted sentiment with actual reviews. By providing a thorough method for sentiment analysis and prediction in social media app evaluations, this work implies that these models can be proactive tools for raising user satisfaction while improving app features in response to expected user input.