Google Play Store Apps Analysis using NLP

Read full text Summarise with VIRAT
Authors: Akshata Vijay Lembhe, Seema Nitin Dokrimare, Rupali Laxman Kamthe, Abhijeet Swami, Arati Y. Dange
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The Google Play Store is one of the world’s largest Android app marketplaces, hosting on the order of two million applications. To help developers interpret the vast volume of user feedback, we analyze a publicly available dataset of 9,146 Google Play apps (from Kaggle, covering dozens of categories). We conduct enhanced exploratory data analysis (EDA) of app metadata and then apply Natural Language Processing to user reviews to classify sentiment. Specifically, we train four supervised classifiers (Naïve Bayes, Random Forest, Logistic Regression, and K-Nearest Neighbors) on the preprocessed review text. We evaluate each model using accuracy, precision, and recall. Our experiments show that logistic regression attains the highest accuracy (around 90%), with balanced precision and recall, while Random Forest and KNN perform comparably (≈87-89%) and Naïve Bayes trails behind (~59%). These updated results provide insights into app market trends: for example, Games and Communication apps dominate install totals, and over 97% of apps are free. Most user reviews are positive (as reported in prior app-review studies), reinforcing that sentiment analysis can help developers understand user preferences and improve app quality.

Fields of Study & Subjects

Sentiment Analysis and Opinion Mining Text and Document Classification Technologies Computational and Text Analysis Methods

Publication details

Journal / source
INTERNATIONAL JOURNAL OF MATHEMATICS AND COMPUTER RESEARCH
Volume
14
Issue
03
Publication date
28 March 2026
Language
EN