Sentiment analysis, market news pipelines, alternative data sources and systematic trading research. Projects here explore natural language processing, alternative features and statistical frameworks that translate market mood into actionable systematic strategies.
Featured Work
MoodSwing Trading: Sentiment-Driven Trading Signals
A real-time sentiment platform that ingests financial news, scores it with an LLM-augmented ensemble, and aggregates the results into a 0–100 MoodScore. Signals are impact-weighted, EWMA-smoothed, and paired with μ/σ forecasts, all exposed through a live dashboard with auditable headline contributions.
PythonspaCyNLTKWebSockets+1 more