I built a live prediction-market analytics platform that processes 300k+ records and delivers real-time insights. Its snapshot-aware pipeline uses PostgreSQL and Python for incremental updates and historical tracking.
I automated daily ETL and analytics with GitHub Actions, then trained and deployed clustering, PCA, and velocity-based models for wallet classification and behavioural prediction. The platform includes a 10-page dashboard on Streamlit Cloud with two live ML inference models.
For a marketplace bounty and contributor intelligence system, I built a Python pipeline with GitHub Actions and Neon PostgreSQL. I also developed web scraping solutions for real estate listings, preparing and standardizing the data for market analysis and forecasting.
