Miheret-Girmachew/fintech-review-analysis

📱🔍 Customer sentiment decoded: Analyzing reviews of top Ethiopian banking apps with NLP & SQLite. Actionable insights for CBE, BOA, and Dashen Bank. 🧠📊 # B5W2: Customer Experience Analytics for Fintech Apps ## Project Overview This project analyzes customer satisfaction with mobile banking apps for three major Ethiopian banks: Commercial Bank of Ethiopia (CBE), Bank of Abyssinia (BOA), and Dashen Bank. The goal is to simulate the role of a Data Analyst at a consulting firm, advising these banks on how to improve their applications based on user feedback. The project involves a full data engineering and analysis pipeline: 1. **Data Collection & Preprocessing**: Combining and cleaning raw review data from CSV files. 2. **Sentiment & Thematic Analysis**: Using NLP (Hugging Face Transformers, spaCy) to extract sentiment and key user themes (e.g., performance, UI/UX). 3. **Database Storage**: Engineering a relational database schema and storing the processed data in a lightweight SQLite database. 4. **Insights & Visualization**: Creating visualizations to identify satisfaction drivers, pain points, and actionable recommendations for the banks. --- ## Project Structure ``` . ├── visuals/ # Contains all generated charts ├── bank_reviews.db # The SQLite database file (often gitignored) ├── boa.reviews.csv # Raw data input ├── cbe.reviews.csv # Raw data input ├── dashen.reviews.csv # Raw data input ├── combine_csv.py # Script to combine and preprocess the raw CSVs ├── analysis.py # Script for sentiment and thematic analysis ├── load_to_db.py # Script to load analyzed data into the SQLite database ├── generate_sql_dump.py # Script to create an SQL dump from the database ├── visualize.py # Script to generate all visualizations ├── create_report.py # Script to compile the final PDF report ├── schema.sql # SQL schema for the SQLite database ├── database_dump.sql # SQL dump file with INSERT statements for all data ├── report.md # Markdown content for the final rep …