AI-powered customer experience analytics platform for Ethiopian fintech apps using NLP, PostgreSQL, automated testing, and interactive dashboards.
# Fintech Review Analytics
## Overview
Fintech Review Analytics is an end-to-end analytics project that analyzes Google Play Store reviews of major Ethiopian banking applications—Commercial Bank of Ethiopia (CBE), Bank of Abyssinia (BOA), and Dashen Bank—to extract actionable insights on customer experience, sentiment, and recurring product issues.
The project transforms unstructured customer feedback into structured business intelligence that helps financial institutions improve mobile banking reliability, enhance customer satisfaction, and prioritize product improvements based on real user experiences.
---
## Business Problem
Digital banking applications receive thousands of customer reviews containing valuable feedback about usability, reliability, and performance. Manually reviewing this feedback is time-consuming and difficult to scale, making it challenging for product teams to identify recurring issues and prioritize improvements.
This project provides an automated analytics pipeline that processes customer reviews, identifies sentiment trends, uncovers common themes, and generates insights that support data-driven product and business decisions.
---
## Solution Overview
The solution implements a complete analytics workflow that:
* Collects customer reviews from Google Play Store
* Cleans and preprocesses review data
* Performs sentiment analysis using a transformer-based NLP model
* Categorizes recurring customer issues into business themes
* Stores processed data in PostgreSQL
* Produces visualizations and business recommendations for decision-makers
---
## Banks Analyzed
* Commercial Bank of Ethiopia (CBE)
* Bank of Abyssinia (BOA)
* Dashen Bank
---
## Dataset
**Source:** Google Play Store
**Total Reviews:** 1,500+
### Fields Collected
* Review Text
* Rating (1–5 Stars)
* Review Date
* Bank Name
* Review Source (Google Play Store)
---
## Tech Stack
* Python 3
* pandas
* NumPy
* scikit-learn
* google-play-scraper
* Hugging Face Tran …