This project leverages data engineering, machine learning, and financial risk analysis to forecast energy prices and assess market risks. I have built an end-to-end pipeline integrating Apache Airflow, PostgreSQL, PyTorch, SQL, and Power BI to analyze energy and weather data from five Spanish cities.
-
ETL Pipeline: Designed a scalable Apache Airflow pipeline to automate data ingestion, transformation, and storage in PostgreSQL.
-
Energy Price Forecasting: Developed an LSTM-based model in PyTorch, achieving 0.0202 RMSE on test data.
-
Real-Time Dashboard: Integrated insights into an interactive Power BI dashboard for visualization and decision-making.
-
Risk Analysis: Calculated 7-day VaR, CVaR, and volatility using SQL for market risk assessment.
- Energy generation and price data (Spain, 5 cities)
- Weather data (temperature, wind speed, humidity, etc.)
- Dataset link: https://www.kaggle.com/datasets/nicholasjhana/energy-consumption-generation-prices-and-weather
- Apache Airflow – ETL pipeline automation
- PostgreSQL – Data storage and management
- PyTorch – LSTM model for time-series forecasting
- SQL – Risk metrics computation (VaR, CVaR, volatility)
- Power BI – Interactive dashboard for insights
- LSTM Model Performance: Achieved 0.0202 RMSE in forecasting energy prices.
- Risk Metrics Computed: 7-day VaR, CVaR, and volatility for market risk assessment.