michaWorku/fintech-review-analytics

AI-Powered Mobile Banking Review Insights. End-to-end data pipeline to extract strategic insights from Ethiopian mobile banking app reviews. # 🔍 AI-Powered Mobile Banking Review Insights > End-to-end data pipeline to extract strategic insights from Ethiopian mobile banking app reviews. ## 🧭 Project Overview This project explores the use of **AI and NLP techniques** to extract meaningful **user insights** from **Google Play reviews** of three major Ethiopian banks: * 💳 Commercial Bank of Ethiopia (CBE) * 🏛️ Bank of Abyssinia (BOA) * 💰 Dashen Bank The goal is to deliver actionable recommendations for **mobile banking app improvement** using real customer sentiment data. ## 🎯 Business Objective Omega Consultancy, a fintech advisory firm, needs: * 🔍 Analysis of app store reviews for major banks * 📊 Strategic insights based on sentiment and user themes * 🗃 Enterprise-compliant storage for structured data * 🛠 Clear, data-driven UX improvement recommendations ## ✅ Tasks & Deliverables | Task | Description | Status | | ------ | ----------------------------------------------------------- | ------ | | Task 1 | Scrape, clean, and preprocess reviews | ✅ Done | | Task 2 | Perform sentiment & thematic analysis | ✅ Done | | Task 3 | Store cleaned data in Oracle DB | ✅ Done | | Task 4 | Extract insights, create visualizations, write final report | ✅ Done | ## 🛠 Tools & Technologies * `Python`, `pandas`, `nltk`, `spacy`, `transformers` * `TextBlob`, `VADER`, `DistilBERT` * `TF-IDF`, `Seaborn`, `Matplotlib`, `WordCloud` * `Oracle XE` (via Docker) for enterprise data storage * `oracledb` Python client for DB operations * Git/GitHub: version control and PR-based workflow ## 🧠 Project Structure ``` 📁 project-root/ │ ├── data/ # Raw and processed review data │ ├── raw/ # Scraped CSVs | ├── clean/ # Cleaned reveiw │ └── processed/ # Cleaned reviews with sentiment & themes │ ├── notebooks/ # Exploratory & production-ready …