Wave-eer/ethiopian_bank_review

# Ethiopian Banks Reviews Scraper & Sentiment Analysis This project collects, cleans, and analyzes customer reviews from the Google Play Store for three major Ethiopian banks: 1. **Commercial Bank of Ethiopia (CBE)** - App ID: `prod.cbe.birr` (CBEBirr Plus) 2. **Bank of Abyssinia (BOA)** - App ID: `com.boa.boaMobileBanking` (BoA Mobile) 3. **Dashen Bank** - App ID: `com.dashen.dashensuperapp` (Dashen Bank Super App) It implements data collection, text preprocessing, a modular NLP pipeline, VADER-based sentiment analysis, TF-IDF thematic keyword extraction, PostgreSQL loading, and visual insights. ## Project Structure ```text ethiopian_banks_reviews/ ├── .github/ │ └── workflows/ │ └── unittests.yml # CI/CD configuration to run unit tests ├── data/ # Scraped and processed CSV data (git-ignored) ├── notebooks/ │ └── images/ # Generated visualization charts (git-ignored/documented) ├── reports/ │ └── insights_recommendations.md # Analysis of satisfaction drivers & recommendations ├── scripts/ │ ├── scrape_reviews.py # Scrapes reviews from Google Play Store │ ├── preprocess_data.py # Cleans reviews (dedupes, drops missing values, formats date) │ ├── sentiment_thematic_analysis.py # Performs VADER sentiment analysis, TF-IDF, and maps themes │ ├── insert_data.py # Inserts preprocessed & analyzed reviews into PostgreSQL │ └── generate_insights.py # Generates sentiment and theme distribution plots ├── src/ │ ├── preprocessing.py # Reusable preprocessing library function │ ├── nlp_pipeline.py # Modular tokenization, stop-word removal, and lemmatization │ └── schema.sql # PostgreSQL schema defining tables and constraints ├── tests/ │ └── test_preprocessing.py # Unit tests for the preprocessing stage ├── .gitignore # Prevents committing datasets, virtual envs, etc. ├── requirements.txt # Project Python dependencies └── RE …