This document analyzes app ratings on the Google Play Store using natural language processing and machine learning techniques. It aims to predict app ratings based on user reviews to identify potential discrepancies between numerical ratings and text reviews. The researchers cleaned and preprocessed a dataset of Google apps and associated reviews scraped from the Play Store. They applied various machine learning classifiers including random forest, decision tree regression, and polynomial regression to predict ratings based on text features extracted from reviews. The random forest model achieved the best performance. The research found that 24.72% of user-defined ratings are inconsistent with the sentiment expressed in associated text reviews. This work demonstrates the potential for using machine learning on app reviews to provide more accurate ratings that better reflect user sentiment.