This project analyzes Google Play Store applications using Data Cleaning, Exploratory Data Analysis (EDA), Inferential Statistics, Feature Engineering, Machine Learning, and an Interactive Streamlit Dashboard.
The objective is to identify the factors that contribute to app success and build a predictive model that estimates an application's success category based on characteristics such as category, pricing strategy, content rating, and app size.
- Analyzed 10,841 apps and 64,295 reviews
- Built 15+ SQL analytical queries
- Identified Game category as largest market segment
- Found free apps significantly outperform paid apps
- Developed ML model and Streamlit dashboard for app success prediction
Developers often struggle to determine whether an app idea is likely to succeed before launch.
This project aims to answer:
- Which categories perform best?
- Do free apps outperform paid apps?
- How do ratings and reviews impact installs?
- Which factors influence app success?
- Can we predict app success before launch?
The project uses two datasets:
Contains:
- App Name
- Category
- Rating
- Reviews
- Size
- Installs
- Type
- Price
- Content Rating
- Genres
- Android Version
- Current Version
Contains:
- User Reviews
- Sentiment
- Sentiment Polarity
- Sentiment Subjectivity
- Python
- SQL (MySQL)
- Pandas
- NumPy
- Matplotlib
- Seaborn
- Scikit-Learn
- SciPy
- Joblib
- Power BI
- Streamlit
- Jupyter Notebook
- VS Code
- Git
- GitHub
The project includes:
- Dataset Shape
- Data Types
- Missing Values
- Duplicate Analysis
- Missing Value Treatment
- Duplicate Removal
- Data Type Conversion
- Install Count Cleaning
- Price Cleaning
- Feature Standardization
- Rating Distribution
- Reviews Distribution
- Installs Distribution
- Price Distribution
- Category Distribution
- Reviews vs Installs
- Rating vs Reviews
- Free vs Paid Apps
- Category vs Ratings
- Correlation Analysis
- Sentiment Analysis
- Category Performance Analysis
Calculated:
- Mean
- Median
- Standard Deviation
- Variance
- Skewness
- Kurtosis
- Quartiles
Objective:
Compare ratings of Free vs Paid Apps
Objective:
Determine whether ratings differ significantly across app categories.
Predict App Success Category using app characteristics.
Success Category:
- Low
- Medium
- High
- Category
- Type
- Content Rating
- Price
- Size
- Logistic Regression
- Linear Regression
- Accuracy Score
- Precision
- Recall
- F1 Score
- Feature Importance
A complete MySQL analytics workflow was implemented before Python analysis.
- Imported 10,841 app records and 64,295 review records using
LOAD DATA LOCAL INFILE - Removed corrupted records
- Handled duplicate apps using
ROW_NUMBER() - Cleaned Installs, Price, and Size columns
- Created an analytics-ready table (
googleplaystore_clean)
- Common Table Expressions (CTEs)
- Multiple CTEs
- Window Functions (
ROW_NUMBER,DENSE_RANK) - Date Functions (
STR_TO_DATE,YEAR) - Joins
- Conditional Aggregation
- Query Optimization
- Sentiment Analysis
- Top Categories by Installs
- Free vs Paid App Analysis
- Content Rating Analysis
- Category Performance Analysis
- Sentiment Analysis by Category
- Top Apps Within Each Category
- App Update Trends
- GAME category generated over 13.4 Billion installs, making it the largest category on the platform.
- Free apps significantly outperformed paid apps in both installs and user engagement.
- Education achieved the highest average rating (4.31), indicating strong user satisfaction.
- Apps rated Everyone dominated the marketplace by total app count.
- Categories such as Video Players, Social, Photography, and Entertainment performed above platform averages.
- Positive review sentiment was observed across most major categories.
- The majority of applications were updated during 2018, highlighting rapid ecosystem growth.
- User engagement follows a winner-takes-most pattern where a small number of apps capture most installs and reviews.
- Prioritize Game, Communication, and Social categories for maximum market reach.
- Adopt a freemium pricing model, as free apps dominate both app count and installs.
- Maintain regular updates to improve discoverability and user retention.
- Optimize application size to improve accessibility across devices.
- Monitor ratings and review sentiment continuously to identify user experience issues early.
An interactive dashboard was developed to visualize insights and allow app success prediction.
link - https://knihrsafpfuwk3g9au8q5e.streamlit.app/
- Dataset Overview
- KPI Metrics
- EDA Dashboard
- Interactive Success Predictor
googleplaystore/
│
├── app.py
├── requirements.txt
├── README.md
│
├── SQL/
│ └── eda.sql
│
├── PowerBI/
| ├── GooglePlayStore_Analytics.pbix
| ├── pb1.png
│ └── pb2.png
│
├── dataset/
│ ├── googleplaystore.csv
│ └── googleplaystore_user_reviews.csv
│
├── cleaned dataset/
│ └── googleplaycleaneddf.xlsx
└── googleplaystore_cleaned.csv
│
├── ML Models/
│ ├── LogisticRegression.ipynb
│ └── LinearRegression.ipynb
│
├── screenshots/
| ├── googleplayimage.png
│ ├── sc1.png
│ ├── sc2.png
│ ├── sc3.png
│ ├── sc4.png
│ └── sc5.png
│
└── googleplayeda.ipynb
This project can help:
- App Developers
- Product Managers
- Marketing Teams
- Data Analysts
understand:
- Factors driving application success
- High-performing categories
- User engagement patterns
- Pricing strategy effectiveness
- Market opportunities
- App launch planning
- SQL Analytics
- Data Cleaning & Wrangling
- Exploratory Data Analysis (EDA)
- Data Visualization
- Descriptive Statistics
- Inferential Statistics
- Hypothesis Testing
- T-Test
- ANOVA
- Correlation Analysis
- Feature Engineering
- Machine Learning
- Logistic Regression
- Linear Regression
- Model Evaluation
- Power BI Dashboard Development
- Streamlit Development
- Business Insights & Data Storytelling
Swapnil Nicolson Dadel
Aspiring Data Analyst passionate about:
- Data Analytics
- Statistics
- Machine Learning
- Business Intelligence
- Power BI
- Python
- SQL
- GenAI
⭐ If you found this project useful, consider giving it a star.
contact - swapnilnicolson.201@gmail.com