Customer Feedback ETL Pipeline
An end-to-end pipeline that extracts, transforms, and loads customer feedback data into a centralized database to generate actionable insights.
Overview
This project implements a robust ETL pipeline that extracts raw customer feedback from CSV files and various APIs, transforms the data by cleaning and enriching it with sentiment analysis (achieving 92% accuracy), and loads the processed data into a centralized PostgreSQL database. The resulting insights are visualized via a custom dashboard for actionable business intelligence.
Key Features
Data Extraction
Gathers customer feedback from CSV files and APIs, automating data ingestion.
Data Transformation
Cleans and processes data with Python and Pandas, applying sentiment analysis using TextBlob.
Data Loading & Visualization
Loads enriched data into PostgreSQL via SQLAlchemy and visualizes results with a custom Streamlit dashboard.
Technical Details
Data Sources
- CSV files from customer surveys
- APIs for real-time feedback data
Software Stack
- Python with Pandas and TextBlob for ETL and sentiment analysis
- Flask for API services
- Supabase and PostgreSQL for data storage
- SQLAlchemy for ORM and database interactions
- Streamlit for dashboard visualizations
What's Next for the Customer Feedback ETL Pipeline
I plan to integrate additional data sources, enhance our sentiment analysis algorithms, and further optimize our data loading processes. Future iterations will include machine learning models for predictive analytics and automated reporting.
