weldie10/mobile-banking-customer-feedback-analysis

A fintech analytics project that scrapes and analyzes Google Play Store reviews for major Ethiopian banks to extract sentiment, themes, and customer insights, supported by a PostgreSQL database and visual reports. # AI Mastery: Mobile Banking Customer Feedback Analysis - Week 0 Challenge Kickstart your AI Mastery with comprehensive customer experience analytics for Ethiopian mobile banking apps. This project focuses on understanding, exploring, and analyzing customer reviews from Google Play Store to identify satisfaction drivers, pain points, and actionable insights for Commercial Bank of Ethiopia (CBE), Bank of Abyssinia (BOA), and Dashen Bank. ## Table of Contents - Project Overview - Business Objective - Dataset Overview - Folder Structure - Setup & Installation - Tasks Completed - Technologies Used - Key Observations ## Project Overview Week 2 challenge for AI Mastery focuses on end-to-end customer feedback analysis: - Data collection, cleaning, and preprocessing - Sentiment and thematic analysis using NLP - PostgreSQL database integration - Cross-bank comparison and visualization - Insights generation with actionable recommendations The project uses real customer review data from Google Play Store for three major Ethiopian banks. ## Business Objective Omega Consultancy seeks to enhance customer retention and satisfaction for Ethiopian banks via data-driven insights from mobile app reviews. Your task is to analyze customer feedback and provide actionable recommendations for identifying satisfaction drivers, addressing pain points, and improving app performance. ## Dataset Overview The dataset contains customer review data from Google Play Store: | Column | Description | |--------|-------------| | `review` | Customer review text content | | `rating` | Star rating (1-5) | | `date` | Review posting date (YYYY-MM-DD) | | `bank` | Bank name (CBE, BOA, Dashen) | | `source` | Data source (Google Play) | | `sentiment_label` | Sentiment classification (positive/negative/neutral) | | `sentiment_score` | Sentiment score (0.00 to 1.00) | | `themes` | Identified themes (semicolon-separated) | **Dataset Statistics:** - Total reviews: 9,571 (697% of 1,200 target) - Missing …