Stack
Next.js
TypeScript
Tailwind
FastAPI
Python
pandas
scikit-learn
◆FRED / OWID
← Back to portfolio
Featured case study
PetroView
- Energy intelligence
- Full-stack
- Forecasting & regimes
Dashboards and pipelines for oil-market signals—structure, risk, volatility, and model-driven views. Built for reproducibility with clear lineage from ingestion through the UI.
Product preview
Live cross-fade from repo screenshots. Swap in assets/petroview-demo.gif when ready.
Impact
PetroView pulls global oil signals into one operational lens: volatility, regimes, and forecasts sit next to narrative context, so decisions aren’t split across spreadsheets. The stack is reproducible from ingest through the dashboard.
Architecture
Frontend: Next.js (App Router), TypeScript, Tailwind, Recharts.
API & modeling: FastAPI, SQLModel, pandas, statsmodels, scikit-learn.
Data: FRED and OWID energy and macro series.
Shipped
- Time series ingestion, normalization, and validation
- Supply, demand, and pricing signal analysis
- Forecasting: naive, regression, ARIMA, and ensembles
- Regime and volatility tracking for live dashboards
- Interactive dashboards and exportable research outputs