Abstract

The Play Store App Review Analysis project aims to explore and analyze customer reviews of Android apps from the Google Play Store to derive actionable insights for developers. With thousands of new applications regularly uploaded on the Play Store, understanding customer demands and factors contributing to app success is crucial. The project investigates relationships among various attributes such as app category, rating, size, pricing, and user reviews.

Project Files Description

• Executable Files:
  • Play Store App Review Analysis: Contains functions for clustering operations.
• Output:
  • Google Colab Notebook: All outputs are visible in the provided Colab notebook.
• Input Files:
  • Play Store Data.csv: Basic details of apps (e.g., reviews, ratings).
  • User Reviews.csv: User reviews and sentiment scores.
• Data Source:
  • Dataset obtained from Almabetter.

Introduction

The project explores a dataset with 10k Play Store applications, analyzing different categories such as family, communication, entertainment, tools, music, and more. Key questions include what makes an app popular, the impact of pricing and size, and trends in user sentiments. The dataset consists of Play Store data and user reviews.

Problem Statements

The project addresses various questions, including:

• Top categories on the Play Store.
• Distribution of free and paid apps.
• Importance of app ratings.
• Popular app categories based on audience.
• Categories with the most installations.
• Analysis of app count by genres.
• Impact of last update on ratings.
• Ratings for paid and free apps.
• Correlation between reviews and ratings.
• Sentiment analysis and its impact on app success.
• Effect of content rating on apps.
• Influence of last update date on ratings.
• Distribution of app updates over the year.
• Distribution of paid and free app updates over the month.

Exploratory Data Analysis (EDA)

Exploratory Data Analysis involves understanding the dataset's structure, cleaning data, and transforming features. The process includes univariate and bivariate analyses, data wrangling, and developing meaningful visualizations.

Steps Involved

1. Problem Statement:
   • Understand the dataset and brainstorm the problem statement.
2. Hypothesis:
   • Develop basic hypotheses based on attributes.
3. Univariate Analysis:
   • Analyze single attributes for patterns and relationships.
4. Bivariate Analysis:
   • Explore cause and relationship between two attributes.
5. Multivariate Analysis:
   • Analyze more than two variables simultaneously.
6. Data Cleaning:
   • Handle errors, duplicates, and NaN values.
7. Testing Hypothesis:
   • Check assumptions for multivariate techniques.

Challenges Faced

Challenges encountered include handling NaN values, merging dataframes, and exploring machine learning opportunities for deeper insights.

Conclusion

The EDA provides valuable insights, such as:

• Majority of apps are free (92%).
• 80% of apps have no age restrictions.
• Competitive categories: Family, Communication.
• Most installs: Game category.
• Most competitive paid app category: Finance.
• Popular app: Facebook.

The analysis guides developers in app development, pricing, and updates. Further exploration and machine learning models are suggested for future work.