ML

Trader Behavior Insights

Advanced trader performance and behavioural analysis across multiple market sentiment regimes.

ML
💹

4

Market Regimes

5+

Performance Metrics

Interactive

Plotly Dashboards

Real-time

Anomaly Flagging

The Problem

Trading desks accumulate vast transaction logs but rarely perform rigorous behavioural analysis. Which traders perform differently in bull vs bear markets? Do certain strategies fail under high-volatility regimes? These questions require multi-regime statistical analysis, not just simple P&L reporting.

The Solution

Built a Python analytics pipeline that segments trading history into market regimes (bull, bear, sideways, high-volatility) using rolling statistics. Computes per-trader metrics — Sharpe ratio, drawdown, win rate, expectancy — within each regime. Visualised with interactive Plotly dashboards revealing hidden behavioural patterns.

Results & Metrics

  • Identifies regime-specific performance breakdowns invisible in aggregate stats
  • Interactive dashboards enable drill-down from portfolio to individual trader
  • Automated anomaly flagging for outlier behaviour
  • Sharpe, Sortino, Calmar, and custom metrics computed per regime
  • Delivered as a self-contained Jupyter + Plotly report

Tech Stack

PythonPandasPlotlyNumPyJupyterscikit-learn