Bilen189/ethiopian-bank-review-analytics

Task 1: Data Collection & Preprocessing Project Overview This project is part of the 10 Academy Artificial Intelligence Mastery Program (Week 2 Challenge). The goal of Task 1 is to collect, clean, and prepare Google Play Store reviews for Ethiopian banking mobile applications for further sentiment and thematic analysis. The banks analyzed include: Commercial Bank of Ethiopia (CBE) Bank of Abyssinia (BOA) Dashen Bank Scraping Methodology Data Source All data was collected from the Google Play Store using the google-play-scraper Python library. Tools Used google-play-scraper → for extracting user reviews pandas → for data manipulation and cleaning numpy → for numerical operations Extraction Process For each banking application: The unique Google Play package ID (app ID) was identified. The reviews() function from google-play-scraper was used to extract user reviews. Reviews were collected in English language (lang='en') and filtered by country (country='et'). A maximum of 2000 reviews per app was requested to ensure sufficient data coverage. Data Fields Collected For each review, the following attributes were extracted: Review ID User name Review text Rating (1–5 stars) Number of thumbs up Review date Bank/app name Date Range Used The dataset includes reviews collected across the full available historical range returned by the Google Play Store API at the time of extraction. Start Date: Earliest available reviews in Google Play Store dataset End Date: Latest available reviews up to the scraping date (May 2026) Note: The dataset is dynamic and reflects the most recent available reviews at the time of scraping. Data Preprocessing Steps The raw dataset was cleaned using the following steps: Column Selection Retained only relevant fields for analysis Renaming Columns Standardized column names for consistency: review_id, review, rating, review_date, bank Missing Value Handling Removed rows with missing review text or ratings Duplicate Removal Removed dupli …