Ruhama-Beshada/omega-fintech-review-analytics-Week-2

A data engineering and NLP project analyzing customer reviews from the Google Play Store for Ethiopian banking apps (CBE, BOA, Dashen). Includes web scraping, sentiment analysis, thematic keyword extraction, PostgreSQL storage, and actionable insights to improve customer experience in fintech. # 10 Academy – AI Mastery Program ## Week 2 Challenge: Customer Experience Analytics for Fintech Apps **Project:** Scraping, Analyzing, and Visualizing Google Play Store Reviews for Ethiopian Banks **Date:** 26 Nov – 02 Dec 2025 --- ## 🎯 Challenge Overview This project simulates a real-world data engineering and analytics task. You act as a **Data Analyst at Omega Consultancy**, supporting three Ethiopian banks: - Commercial Bank of Ethiopia (CBE) - Bank of Abyssinia (BOA) - Dashen Bank **Objective:** Scrape user reviews from Google Play, preprocess text, analyze sentiment and themes, visualize insights, and deliver actionable recommendations to improve mobile app performance and user satisfaction. **Focus Areas:** - **User Retention:** Identify common complaints (e.g., slow transfers) and patterns in ratings. - **Feature Enhancement:** Extract desired features and improvement requests from reviews. - **Support Efficiency:** Cluster complaints to guide support strategies, including AI chatbot integration. --- ## 🗂 Dataset Overview - **Source:** Google Play Store - **Fields Collected:** - `review_text` → User feedback - `rating` → 1–5 stars - `review_date` → Date of review - `bank_name` / `app_name` → Bank identifier - `source` → Google Play - **Minimum Reviews:** 400 per bank (1,200 total) --- ## 🚀 Project Tasks & Deliverables ### **Task 1: Data Collection & Preprocessing** - **Web Scraping:** Use `google-play-scraper` or BeautifulSoup to collect reviews. - **Preprocessing:** Remove duplicates, normalize dates, clean text (remove emojis, URLs, stopwords). - **Deliverables:** - Cleaned CSV with columns: `review`, `rating`, `date`, `bank`, `source` - GitHub repo with organized commits in **task-1** branch **KPIs:** ≥1,200 reviews, 1,000 reviews, schema documented in README --- ### **Task 4: Insights & Recommendations** - **Analysis:** Identify key drivers (e.g., speed) and pain points (e.g., crashes) per bank - **Visualization:** 3–5 plots (sentiment …